Full text of "Romanising Oriental Gods"

tCO-IOKA» 



tolllfclllttlllEfJ 



Oriental Gods 

Myih, Salvation and Ethics 

in the Cults of Cyliefc. 

Jsfertnd Mithras 



|.n vii H.VA1 



Tnmtbtor irut odi (or 



Romanising Oriental Gods 



PDF processed with CutePDF evaluation edition www.CutePDF.com 



Religions in the 
Graeco-Roman World 



Editors 

H.S. Versnel 

D. Frankfurter 

J. Hahn 



VOLUME 165 



Romanising Oriental Gods 

Myth, Salvation and Ethics in the Cults of 
Cybele, Isis and Mithras 



by 
Jaime Alvar 

Translator and editor 

Richard Gordon 









BRILL 



LEIDEN • BOSTON 

2008 



The publication of this work has been made possible through a subsidy received 
from the Directorate General for Books, Archives and Libraries of the Spanish 
Ministry of Culture. 

This book is printed on acid-free paper. 

Library of Congress Cataloging-in-Publication Data 

Alvar, Jaime. 

Romanising oriental Gods : myth, salvation, and ethics in the cults of Cybele, Isis, and 
Mithras / by Jaime Alvar ; translator and editor Richard Gordon, 
p. cm. — (Religions in the Graeco-Roman world ; 165) 

Includes bibliographical references and index. 

ISBN 978-90-04-13293-1 (hardback : alk. paper) 1. Cybele (Goddess)— Cult. 
2. Attis (God) — Cult. 3. Isis (Egyptian deity) — Cult. 4. Serapis (Egyptian deity) — Cult. 
5. Mithras (Zoroastrian deity) — Cult. 6. Rome — Religion. I. Alvar Ezquerra, Jaime. 
I. Title. II. Series. 

BL820.C8R66 2008 
200.937— dc22 

2008015353 



ISSN 0927-7633 

ISBN 978 90 04 13293 1 

Copyright 2008 by Koninklijke Brill NY Leiden, The Netherlands. 
Koninklijke Brill NV incorporates the imprints Brill, Hotei Publishing, 
IDC Publishers, Martinus Nijhoff Publishers and VSP 

All rights reserved. No part of this publication may be reproduced, translated, 
stored in a retrieval system, or transmitted in any form or by any means, electronic, 
mechanical, photocopying, recording or otherwise, without prior written permission 
from the publisher. 

Authorization to photocopy items for internal or personal use is granted by 
Koninklijke Brill NV provided that the appropriate fees are paid directly to 
The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, 
Danvers, MA 01923, USA. 
Fees are subject to change. 

PRINTED IN THE NETHERLANDS 



$6ey^o|iai oi<; 6e|ii<; ecru, 0-upa<; 8e e7ii0ea6e 
fiepriXov. 

'Orpheus' as cited by Aristobulus ap. Eusebius, 

Praeparatio evangelica 13.12.5 (vol. XII— XIII p. 312 

des Places) = 247 Kern = 378 F Bernabe 

Quippe cum transactis vitae temporibus iam in ipso finitae 
lucis limine constitutes, quis tamen tuto possint magna 
religionis committi silentia, numen deae soleat elicere et sua 
providentia quodam modo renatos ad novae reponere rursus 
salutis curricula. 

The chief priest, in Apuleius, Met. 1 1.21 

I will not say that the tragic world-view was every- 
where completely destroyed by this intruding un- 
Dionysian spirit: we only know that it had to flee from 
art into the underworld as it were, in the degenerate 
form of a secret cult. 

Fr. Nietzsche, The Birth of Tragedy (1884) §17 



To Pablo and Irene Alvar Rozas, 

whose lives have given mine a new dimension 



CONTENTS 

Foreword ix 

Translator's Note xi 

Abbreviations xiii 

List of Texts-Figures and Plates xvii 

Introduction 1 

Chapter One Religion, Cult and Mystery 17 

Chapter Two Systems of Belief 25 

1. Cosmic Order and the Nature of the Divine 33 

a. The Egyptian Cults 39 

i. The Myth of Isis and Osiris 39 

ii. Serapis 52 

b. The Myth of Cybele and Attis 63 

c. Mithras 74 

2. Humankind in the World 106 

3. The World Beyond 122 

Chapter Three Systems of Value 143 

1. Between Utopia and Reality 154 

2. Ethics in the Phrygian Cults 165 

3. Isiac Ethics 177 

4. Moral Values in Mithraism 192 

Chapter Four The Ritual Systems 205 

1. Religion and Ritual 206 

2. Rituals in the Mysteries 211 

a. Initiation 217 

b. Sacrifice 221 

c. Commensality 227 

d. Prayer 231 

3. Rituals in the Phrygian Cults 240 

a. Introduction 240 

b. Emasculation 246 



Vlll CONTENTS 

c. The taurobolium/criobolium 261 

d. Initiation 276 

e. The Megalensia and the March Festival of Attis 282 

4. Rituals in the Egyptian Cults 293 

a. Festivals 296 

b. Cultic Practice 305 

i. Sacrifice and Votives 313 

ii. Prayer, Healing and Incubation 318 

c. Initiation 336 

5. Cultic Practice in Mithraism 344 

a. The Ritual Function of the Mithraeum 349 

b. Initiation and the Initiatory Grades 364 

Chapter Five The Oriental Cults and Christianity 383 

1. The Problems 384 

2. The Sub-system of Belief 393 

3. The Sub-system of Ethics 401 

4. The Sub-system of Ritual 405 

5. From Reverse Borrowing to 'Commensality' 417 

Bibliography 423 

Plates 445 

General Index 461 

Index of Inscriptions, Papyri and Monuments 474 

Index of Passages 481 



FOREWORD 

This book has been a long while in the making. During that time I have 
been the beneficiary of generous support, for which it is an inadequate 
repayment. The Direccion General de Investigation Cientifica y Tecnica 
has supported my group of researchers with grants for specific projects 
continuously since 1990. This assistance has been essential both in the 
process of actually writing the book and for the doctoral theses that 
grew up around it. Other institutions have also contributed financial 
help towards its completion. The CSIC gave me one of its bursaries 
as part of a programme of exchanges with the British Academy. A 
grant from the Comunidad Autonoma de Madrid allowed me to spend 
several months at the Institute of Classical Studies in London and 
in Cambridge in 1992. Thanks to the generosity of the Universidad 
Complutense and the Ministerio de Education y Ciencia I was able 
to spend time at the Centre d'Histoire Ancienne at Besancon in 1993, 
and at the Ecole Normale Superieure in the rue d'Ulm in 1994. The 
Junta de Andalucia financed my visits to the Faculty of Classics in 
Cambridge over the years I was a professor in Andalusia. However the 
decisive impulse to finish the book was given by a sabbatical granted 
by the University of Huelva for the academic year 1999-2000, which 
I was able to spend at the Classics Faculty in Cambridge. I have been 
made very welcome at the Library of the DAI in Madrid, and two 
excellent libraries of the Universidad Complutense, the well-stocked 
Biblioteca de Clasicas and the amazing Biblioteca de Humanidades, 
have likewise been of great help. 

Thanks to all this institutional generosity I have been lucky enough 
to make friends with a number of fine historians. I must mention 
here the unforgettable Pierre Leveque, and Monique Clavel-Leveque, 
at Besancon, and Marguerite Garrido-Hory Jacques Annequin and 
Antonio Gonzales, whom I am proud to call my friends. At Cambridge 
Keith Hopkins opened many doors to me; his death in 2004 has 
caused me great sadness. It was there too that I met Richard Gordon, 
who helped to re-direct me when I was at a loss and who has been 
a source of constant stimulation thereafter. Not only did he patiently 
correct the manuscript of the first edition in Spanish, but he agreed to 
write a Foreword to that version; and has now compounded his help 



X FOREWORD 

by undertaking to translate the entire book into English and subject it 
to critical scrutiny. His generous updating of the discussion of many 
topics and addition of much new bibliography have made this English 
version not a mere translation of the old book, but a truly new edition. 
He knows how very grateful I am to him. I would also like to express 
my gratitude to Henk Versnel and his colleagues on the editorial board 
of the series Religions in the Graeco-Roman World for kindly agreeing 
to publish the English version as the first contribution to that series on 
the central topic of its predecessor, MJ. Vermaseren's EPROER. 

I am glad here also to be able to acknowledge my debts to all the 
colleagues at my various universities, the Universidad Complutense in 
Madrid, at Huelva, Pablo de Olavide (Seville) and Carlos III, again in 
Madrid, who have contributed so much to the formation of my think- 
ing and to the arguments of this book. 

My final thanks go to Pilar Rozas, who made it possible for me over 
many years to devote so much of my time to research. 

December 2007 Universidad Carlos III de Madrid 



TRANSLATOR'S NOTE 

Surprising as it may seem, this book was originally intended as an intro- 
ductory text for Spanish university students. To this end, the original 
version was equipped with various glossaries and aids that have been 
omitted here. The present version is directed rather to the academic 
community. Since it is not intended as in any sense an external his- 
tory of these cults, the distribution-maps have been omitted; specialists 
can consult the perfectly adequate maps in Vermaseren's CCCA and 
CIMRM, and L. Bricault's excellent Isiac Atlas (Bricault 2001), which 
sets new standards for such cartographic enterprises. There are many 
other changes too. My main aim has been to turn Alvar's Spanish into 
straightforward academic English. This is emphatically a version and 
not a literal translation. Moreover, I have suggested many changes to the 
argument, called attention to new work, advised omissions, and generally 
intervened in the text so as to fit the work for a new academic audience. 
Moreover some passages in the original appear to a non- Spaniard to be 
tilting at windmills. Even when these have been eased out, there may still 
be some oddities in tone (the undertaking to translate the book was an 
act of friendship, not necessarily an endorsement of the content). Alvar 
too has himself contributed a number of alterations and additions, so 
that the translation amounts to a thoroughly revised edition. 

As far as practicable and illuminating, citations are now given in the 
original Greek or Latin, sometimes (when straightforwardly compre- 
hensible) untranslated. The rather odd dual system of footnotes and 
endnotes in the original has been replaced by a unified system of notes 
at the foot of the page, so that it will be difficult to locate any given 
footnote in the original; there are also, especially in Chap. 4, many 
new footnotes. I have retained Alvar's practice of placing epigraphs 
before each section, but where possible have returned them to their 
original languages. The cumbrous practice in the original of citing 
bibliographic entries in full both in the notes and in a consolidated 
bibliography (though there were many exceptions) has been replaced by 
the system employed by RGRW, whereby the consolidated bibliography 
contains only items cited more than once, with everything else either 
to be found in the list of abbreviations in the prelims or cited in full 
in the footnotes. The consolidated bibliography can therefore be used 



Xll TRANSLATOR'S NOTE 

neither as an accurate judge of what Alvar has read nor what he has 
not: it is simply a list of items that meet a technical requirement. I have 
however not been able to bring myself to employ the stipulated forms 
BCE and CE for the time-honoured BC and AD. 

In view of its target audience, the original edition had 7 7 rather poor- 
quality illustrations. Since it is Brill's policy only to publish illustrations 
required by the argument, the number has been greatly reduced, and 
they are mostly different from those of the original. The new illustra- 
tions are mainly Mithraic: the argument in the sections devoted to Isis 
and the Mater Magna almost never depends upon images; in the case 
of Mithras, where archaeological evidence is anyway of much greater 
importance, it often does. We are very grateful to all the museums, other 
organisations, and individuals for their kind co-operation. The Spanish 
Ministry of Education handsomely provided a substantial grant-in-aid 
for the translation; and has been very forgiving of the delays in ready- 
ing the text for the press. Much of the considerable labour involved 
in preparing this new edition was carried out at two institutions, the 
Getty Villa in Pacific Palisades, Los Angeles, and the Saxo Institute of 
the University of Copenhagen. To colleagues and staff at both I extend 
my grateful thanks. 

January 2008 Universitat Erfurt 



ABBREVIATIONS 

Abbreviations (acronyms) for some frequently-cited books, collective 
works and series are listed below. The epigraphic abbreviations not 
listed here will be found in F. Berard et al., Guide de I'epigraphiste 3 (Paris 
2000) with the downloadable supplements. Abbreviations employed in 
the bibliography for specialist journals mainly conform to those used 
in L'Annee Philologique, sometimes expanded for ease of comprehension, 
so that they need not be repeated here. 



AE 
ANRW 



Atlas 



Ausstellung Liebighaus: 



Barrington 
BEFAR 

BEHE 

CAH 

CCCA 

Chantraine 

CIL 

Collezioni egizie 

CPG 



L'Annee epigraphique 

H. Temporini and W. Haase (eds.), Aufstieg und 
Niedergang der romischen Welt. Geschichte und Kultur 
Roms im Spiegel der neueren Forschung (Berlin and 
New York 19 70-) 

L. Bricault, Atlas de la diffusion des cultes isiaques (IV 
siecle av. J.-C.-IV e siecle apt J.-C.) (Paris 2001) 
H. Beck, P. Bol, M. Buckling (sub cura), Agypten, 
Griechenland, Rom: Abwehr und Beruhrung Stadelsches 
Kunstinstitut und Stadtische Galerie (Liebighaus, Frank- 
furt a.M.), Katalog der Austellung 26. Nov. 2005-26. 
Feb. 2006 (Frankfurt a.M. and Tubingen 2005) 
R.J.A. Talbert (ed.), Barrington Atlas of the Greek 
and Roman World (Princeton 2000) 
Bibliotheque des Ecoles francaises d'Athenes et 
de Rome 

Bibliotheque de l'Ecole des Hautes Etudes 
The Cambridge Ancient History 
MJ. Vermaseren, Corpus cultus Cybelae Attidisque. 
EPROER 50/1-7 (Leyden 1977-89) 
P. Chantraine, Dictionnaire etymologique de la langue 
grecque: Histoire des mots. 4 vols. (Paris 1968-80) 
Corpus inscriptionum latinarum (Berlin 1863—) 
F Manera and C. Mazza, Le collezioni egizie del 
Museo Nazionale Romano (Rome 2001) 
E.L. Leutsch and EG. Schneidewin (eds.), Corpus 
paroemiographorum graecorum. 2 vols. (Gottingen 
1839-51) 



XIV 



ABBREVIATIONS 



CPL R. Cavenaile (ed.), Corpus Papyrorum Latinarum (Wies- 

baden 1958) 

D-K H. Diels and W. Kranz, Die Fragmente der Vorsokratiker. 

Griechisch und Deutsch 10 (Berlin 1960) 

DNP H. Cancik and H. Schneider (eds.), Der Neue Pauly. 16 

vols, in 19 (Stuttgart and Weimar 1996-2003) 

DTM J. Blansdorf (ed.) 2008. Forschungen zum Mainzer 

Isis- und Mater-Magna-Heiligtum, 1: Die Defixionum 
tabellae des Mainzer Isis- und Mater-Magna-Heilig- 
tums (DTM). Mainzer Archaologische Schriften, 1 
(Mainz). 

Electronic Journal of Mithraic Studies: <www.clas.canter- 
bury.ac.nz/ejms> 

Epigraphische Datenbank Heidelberg: http://www. 
uni-heidelberg de/institute/sonst/adw/edh 
Etudes preliminaries aux religions orientales dans 
l'Empire romain, series edited by M.J. Vermaseren. 
113 vols, in 145 (150) (Leyden 1961-1990) 
A. Ernout and A. Meillet, Dictionnaire etymologique de la 
langue latine. Histoire des mots 4 (Paris 1967) 
E Jacoby Die Fragmente der griechischen Historiker (Leyden 
1923—58, cont'd by a variety of other scholars from 
1994-) 

H. Beck and U. Walter (eds.), Diefruhen rbmischen His- 
toriker (Darmstadt 2001—4) 

Garland A.S.E Gow and D.L. Page (eds., comm.), The Greek 

Anthology: Garland of Philip and some Contemporary Poems 
(Cambridge 1968) 

HAaW A.-M. Wittke, E. Ohlshausen and R. Szydlak, Histo- 

rischer Atlas der antiken Welt. DNP Supplement 3 (Stutt- 
gart and Weimar 2007) 

Hellenistic Epigrams A.S.E Gow and D.L. Page (eds., comm.), The Greek 
Anthology: Hellenistic Epigrams (Cambridge 1965) 

Helbig 4 W Helbig, Ftihrer durch die offentlichen Sammlungen klas- 

sischer Altertiimer in Rom* 4 vols. (Tubingen 1963—72) 

IAH EP Bremer, lurisprudentiae Antehadrianae quae supersunt. 



EJMS 

EpDatHeid 

EPROER 

Ernout-Meillet 
FGrH 

FRH 



IESS 
IG 



3 vols, in 2 (Leipzig 1896-1901) 
International Encyclopedia of the Social Sciences 
Inscriptiones graecae (Berlin 1903—) 



ABBREVIATIONS 



ILAquit Inscriptions Mines d'Aquitaine (Bordeaux 1994—) 

ILS H. Dessau, Inscriptiones latinae selectae. 3 vols, in 5 (Berlin 

1892-1916) 
Inscrlt Inscriptiones Italiae 

Kern OF O. Kern, Orphicorum Fragmenta (Berlin 1922) 

LAg W. Helck, E. Otto and W. Westendorf (eds.), Lexikon 

der Agyptologie. 6 vols. (Stuttgart 1975-86) 
LIMC H.C. Ackermann andJ.R. Gisler (eds.), Lexicon iconogra- 

phicum mythologiae classicae. 8 vols, in 16 (Zurich 1981— 

1999) 
LSCG E Sokolowski, Lois sacrees des cites grecques. Travaux et 

memoires de l'Ecole fr. d'Athenes 9 (Paris 1969) 
LSJ 9 H.G. Liddell, R. Scott, H. Stuart Jones and R. McKen- 

zie (eds.), A Greek-English Lexicon with a revised Supplement 

(P.G.W. Glare ed.) (Oxford 1996) 
LTUR M. Steinby (ed.), Lexicon Topographicum Urbis Romae. 6 

vols. (Rome 1993-2000) 
Migne, PG J.P. Migne (ed.), Patrologia graeca. 161 vols. (Paris 1857-66 

and frequently reprinted) 
Migne, PL J.P. Migne (ed.), Patrologia latina. 221 vols. (Paris 1844-64 

and frequently reprinted) 
Mostra hide EA. Arslan (ed.), hide: II mito, il mistero, la magia. Catalogo 

delta Mostra, Milano, Palazzo Reale, 22febb.—l giugno 1997 

(Milan 1997) 
Nilsson, GGR M.P Nilsson, Geschichte der griechischen Religion (Munich: 

vol. I s , 1967; vol. 2 2 , 1961) 
OGIS W. Dittenberger, Orientis graecae inscriptiones selectae, 2 vols. 

(Leipzig 1903-05) 
OLD P.G.W. Glare (ed.), Oxford Latin Dictionary (Oxford 1968 

82) 
PAwB Potsdamer Altertumswissenschaftliche Beitrage 

PGrMag K. Preisendanz (ed.), Papyri Graecae Magicae (ed. 2 by 

A. Henrichs). 2 vols. (Stuttgart, 1973-74; ed. 1 Leipzig 

and Berlin. 3 vols.: 1928; 1931; [1941]) 
RE [PaulysJRealencyclopddie der classischen Altertumswissenschaft, 

neue Bearbeitung von G. Wissowa ... 84 vols. (Stuttgart, 

Weimar, then Munich 1894-1980) 
RfAC Reallexikon filr Antike und Christentum (Stuttgart, 1950—) 

RGG* H.D. Betz et al. (eds.), Religion in Geschichte und Gegenwart 4 

(Tubingen 1998-2007) 



XVI 



ABBREVIATIONS 



RIC H. Mattingly, E.A. Sydenham et al. (eds.), The Roman Imperial 

Coinage (London 1923-81, reprinting 2001—) 
RICIS L. Bricault, Recueil des inscriptions concernant les cultes isiaques. 

3 vols. Memoires de l'Academie des Inscriptions et Belles- 

Lettres 31 (Paris 2005) 
RGRW Religions in the Graeco-Roman World, edited by H.S. Vers- 

nel, R. van den Broek et al., numbered from the last of the 

EPRO series, 114 (1992) (Leyden 1 992-) 
RGW Religionsgeschichtliche Versuche und Vorarbeiten, 1 (1903)— 

27 (1939); 28 (1969)— (formerly GieBen, now Berlin) 
RRC M.H. Crawford, Roman Republican Coinage. 2 vols. (Cambridge 

1974) 
SB F. Preisigke, F. Bilabel, E. Kiessling et al. (eds.), Sammelbuch 

griechischer Urkunden aus Agypten (Berlin 19 15-) 
SIRIS L. Vidman, Sylloge inscriptionum religionis Isiacae et Sarapiacae. 

RGW 28 (Berlin 1969) 
Syll. W. Dittenberger, Sylloge Inscriptionum Graecarum. 3 4 vols. (Leipzig 

1915-24; repr. Hildesheim 1982) 
ThesCRA J.C. Baity et al. (eds.), Thesaurus cultus et rituum antiquorum. 6 vols. 

(Los Angeles 2004-2006) 
TMMM F Cumont, Textes et monuments figures relatifs aux Mysteres de 

Mithra. 2 vols: 1: Introduction; 2: Textes et monuments (Brussels 

1896-99) 
Totti M. Totti, Ausgewdhlte Texte der Isis- und Sarapis-Religion (Hilde- 

sheim 1985) 
TRF O. Ribbeck, Tragicorum Romanorum Fragmenta 3 (Leipzig 

1897). 
TWJVT G. Kittel and G. Friedrich (eds.), Theologisches Worterbuch zum 

JVeuen Testament (Berlin and Stuttgart 1933-79) 
V M.J. Vermaseren, Corpus inscriptionum et monumentorum religionis 

Mithriacae. 2 vols. (The Hague 1956-60) 
Walde-Hofmann A. Walde, Lateinisches etymologisches Worterbuch 3 (ed. 

J.B. Hofmann). 3 vols. (Heidelberg 1938-56) 



LIST OF TEXT-FIGURES AND PLATES 

Text-figures 

1. Plan of the Ptolemaic wall-foundations of the Serapeum- 
complex at Alexandria. Drawn by J. McKenzie. From 
McKenzie et al. 2004, 82 fig. 6. Courtesy of the Roman 
Society, London 55 

2. The site of the Serapeum before the monumentalisation 
by Ptolemy III Euergetes, showing the tunnel linking the 
T-shaped Building and the South Building. Drawn by 

J. McKenzie. From McKenzie et al. 2004, 83 fig. 7. 

Courtesy of the Roman Society, London 57 

3. The ecliptic and zodiac from Taurus to Scorpius, with 
the southern paranatellonta of the summer quadrant 

(c. AD 100). Courtesy of R.L. Beck 96 

4. The 'Entry of the Miles' group on the Mainz 
SchlangengefaB, c. AD 120—40. Courtesy of 
Landesarchaologie Mainz 347 

5. The 'Test of Bravery' group on the Mainz SchlangengefaB, 

c. AD 120-40. Courtesy of Landesarchaologie Mainz 372 

Plates 

Frontispiece: 

Cybele enthroned, with a portrait head. Roman, about AD 50. Prov- 
enance: unknown. The J. Paul Getty Museum, Villa Collection, inv. 
no. 57.AA.19. Photo: Museum. 

1. Reclining solar Attis dedicated by C. Cartilius Euplus, mid-IP. From 
the Attideum of the Metroac complex at Ostia. VatMus MGP inv. 
no. 10785. Photo: Courtesy of the Vatican Museums. 

2. Enthroned Serapis, with staff and Cerberus (so-called Bryaxis type). 
Roman, IP. From the Macellum of Puteoli (Pozzuoli) = Malaise 
1972b, 288 no. 23. Courtesy of Museo Archeologico Nazionale di 
Napoli. 

3. Wooden panel with painted image of Serapis wearing the kalathos 
(part of a triptych). Graeco-Egyptian, IIP. Provenance: unknown. 
The J. Paul Getty Museum, Villa Collection, inv. no. 74.AP21. 
Photo: Museum. 



XV111 LIST OF TEXT-FIGURES AND PLATES 

4. Bronze statuette of Isis-Fortuna (h. 19 cm) wearing a complex 
head-dress (crown of Hathor) consisting of ears of grain, cow's 
horns, solar disc and twin feathers, and holding the cornucopiae 
and ship's steering-oar. Date: imperial. Provenance: unknown. The 
J. Paul Getty Museum, Villa Collection, inv. no. 7 LAB. 180. Photo: 
Museum. 

5. Silver lanx from Parabiago, nr. Milan: Mater Magna and Attis in 
lion-car, with cosmic symbols. Date: IV P . Photo: Soprintendenza 
per i Beni Archeologici per la Lombardia. 

6. Pilaster-relief from Osterburken (Bauland, w. of Heidelberg), found 
in 1861. Date: c. AD 200. Photo of copy in Museum fur Abgiisse 
klassischer Bildwerke, Munich. 

7. Mithras hauling the captured bull into the cave (Transitus dei). From 
Mithraeum I, Poetovio (Spodnja Hajdina). Pokrainski Muzej, Ptuj. 
inv. no. RL 142. Late IP. Courtesy of the Museum. 

8. Damaged Lion-headed god, from the cachette discovered in 1902 
on Cerro de St. Albin outside the ancient walls of Augusta Emerita 
(Merida, Spain). Museo Nacional de Arte Romano de Merida, inv. 
no. 87. Mid-II p . Courtesy of the Museum. 

9. Jupiter destroying the Giants. From Osterburken relief, left pilaster, 
fifth scene from bottom. Photo of copy in Museum fur Abgiisse 
klassischer Bildwerke, Munich. 

10. Mithras born from rock, supported by Cautes and Cautopates, 
with Saturnus sleeping above. Mithras holds the sword in his right 
hand, a torch in his left. From Mithraeum III, Poetovio. Pokrainski 
Muzej, Ptuj. inv. no. RL 296. III 1 '. Courtesy of the Museum. 

1 1. Limestone relief from the Mithraeum of Absalmos (0.66 x 0.49 x 
0.10m). Provenance: unknown, but apparently from Roman Syria. 
Inscr.: AE 1999: 1675 = BulUp. 2001: 481. Probably IIP. Israel 
Museum Collection, inv. no. 97.95.19. Courtesy of the Museum. 

12. Altar showing Mithras carrying a lighted torch and riding the bull. 
Probably from Apulum (Alba Iulia, Romania). Brukenthal National 
Museum, Sibiu, inv. no. 7274. Courtesy of the Museum. 

13. Mithras creating a life-giving spring out of the bare rock by shooting 
an arrow. Altar of Flavius Aper, vir egregius etpraepositus. Mithraeum 
III, Poetovio. Mid-III p . Pokrainski Muzej, Ptuj. inv. no. RL 293. 
Courtesy of the Museum. 



LIST OF TEXT-FIGURES AND PLATES XIX 

14. The establishment of a natural cycle by Mithras: condensed allu- 
sions to the 'water-miracle' (bow and quiver) and the tauroctony 
(sword). Left lateral face of the same altar as in PI. 13. Courtesy 
of Pokrainski Muzej, Ptuj. 

15. Reverse of Mithraic relief from Konjic(a), Bosnia-Hercegovina: 
banquet of Mithras and Sol, with torchbearers and representa- 
tive Corax and Leo. IV P . Copy in former school-museum, Sarajevo. 
Photo: J.R. Hinnells. 

16. Complex relief of M. Aurelius Euthices from Apulum (Alba Iulia, 
Romania). Brukenthal National Museum, Sibiu, inv. no. 7162. 
Photo: Museum. 

17. The Twelve Gods. Osterburken relief, crown of cave-arch. Photo 
of copy in Museum fur Abgiisse klassischer Bildwerke, Munich. 

18. Dexiosis of Mithras and Sol. Osterburken relief, right pilaster, third 
scene from bottom. Photo of copy in Museum fur Abgiisse klas- 
sischer Bildwerke, Munich. 

19. Mithras and Sol roasting meat on a spit over an altar, with the raven 
(= Corax?) coming to eat. Note bull's haunch on the ground. Altar 
of Flavius Aper, vir egregius et praepositus. Mithraeum III, Poetovio. 
Mid-III p . Pokrainski Muzej, Ptuj. inv. no. RL 293. Courtesy of the 
Museum. 

20. Relief of dying Attis leaning against the pine-tree. Note the smaller 
tree behind. Glanum, Musee archeologique, St. Remy de Provence, 
inv. no. 1186. Courtesy of Centre des Monuments Nationaux, 
Hotel de Sade a St. Remy de Provence. 

21. Idealised procession in honour of Cybele: fresco from Via delT- 
Abbondanza, Pompeii. Courtesy of Museo Archeologico Nazionale 
di Napoli. 

22. Relief panel showing four Isiac priests in procession. The hindmost 
priestess carries a simpulum for scattering Nile-water during the pro- 
cession. Formerly in the Coll. Mattei = Helbig 4 1: 388f. no. 1491. 
Perhaps from Rome itself. Courtesy of the Vatican Museums. 

23. Fresco depicting an Ethiopian performing a sacred dance on the 
steps of an Isiac temple decorated with palm branches and a crown. 
Herculaneum, 0.80 x 0.87m. Courtesy of Museo Archeologico 
Nazionale di Napoli. 

24. Beginning of an Isiac procession: a gilt hydria filled with Nile water 
is brought out of the temple by the 7tpo(pr|Tr|<;, while the worshippers 
waiting to take part sing and rattle their sistra. Herculaneum, 0.80 x 
0.85m. Courtesy of Museo Archeologico Nazionale di Napoli. 



XX LIST OF TEXT-FIGURES AND PLATES 

25. General view of the Mithraeum at S. Maria Capua Vetere. 

26. View of Mithraeum I, Poetovio (Spodnja Hajdina), looking towards 
the cult-niche, with the votive altars and other furniture. Photo: 
Pokrainski Muzej, Ptuj. 

27. Relief from Augusta Emerita, from a private house in town (Calle 
S. Francisco, no. 2). Three banqueters at a table, flanked by two 
torchbearers, and a servant with food. On the extreme left, prob- 
ably Mithras' birth. Museo Nacional de Arte Romano de Merida, 
inv. no. 127. Courtesy of the Museum. 

28. Floor mosaic of the mitreo di Felicissimo, Ostia, Reg.V ins. 9, 
looking towards the cult niche (i.e. from the symbols for Corax to 
those of Pater). Mid-III p . Photo: R. Gordon. 

29. The 'Initiation' or 'Obeisance' of Sol. Osterburken relief, right 
pilaster, fourth scene from bottom. Photo of copy in Museum fur 
Abgiisse klassischer Bildwerke, Munich. 

30. Banquet of Mithras and Sol. Terra sigillata dish found in the Skt. 
Matthias Roman cemetery, Trier, in 1905 (diam. 0.175m). IIP. 
Rheinisches Landesmuseum, Trier, inv. no. 05.228. Courtesy of 
the Museum. 



INTRODUCTION 

Pour faire l'histoire d'une religion, il est 
necessaire, premierement, d'y avoir cru (sans cela, 
on ne saurait comprendre par quoi elle a charme et 
satisfait la conscience humaine); en second 
lieu, de n'y plus croire d'une maniere absolue; car 
la loi absolue est incompatible avec l'histoire sincere. 

Ernest Renan 
Vie de Jesus, Introduction 

This book aims to contribute to a fuller understanding of the social and 
ideological reality of the Roman Empire. Its topic is three — and only 
three — of the cults known since Franz Cumont's famous lectures of 
1905—6 (Cumont 1929) as the oriental religions of the Roman Empire, 
but more correctly as graeco-oriental cults: those of Mater Magna 
(Cybele), Isis and Serapis, and Mithras. Its subject is the reception, 
transformation and socio-religious roles of these cults in the complex 
culture of the Roman Empire, and their relation to emergent Chris- 
tianity Its aim is to reassert the validity of the traditional practice of 
viewing these cults as a typological unity, and, against current trends, 
to re-emphasise their significance in the religious history of at any rate 
the Latin-speaking areas of the empire. 

First, a word about terminology. I concede that both terms of the 
traditional expression 'oriental religions' are unsatisfactory. On the first, 
'oriental', no one can today fail to be aware of the ideological load- 
ing of the concept, from the time of the Persian Wars to the colonial 
aftermath of the collapse of the Ottoman Empire in 1918. ' No one can 
now read Cumont's utterances about 'les Levantins', or his descriptions 
of Dio Chrysostom or Lucian as 'orientaux', or his talk of Tinvasion 
des cultes orientaux' and 'mysteres barbares', without misgivings. 2 He 



1 T. Hentsch, L'Orient imaginaire: La vision politique occidentale de I'Est mediterraneen (Paris 
1988) is indispensable; in the context of religion, Amir-Moezzi 2000. 

2 Resp. Cumont 1929, 4; 6; 16, 38. On Cumont's constructions, which betray 
considerable muddle and indecision vis-a-vis the contrasting views of his immediate 
predecessors in France, esp. Ernest Renan and Victor Duruy, see J.-M. Pailler, Les 
religions orientales selon Franz Cumont: une creation continu, MEFR1M 111 (1999) 
635—46; C. Bonnet, Les "religions orientales" au laboratoire de l'Hellenisme, 2: 
F. Cumont, in Bonnet and Bendlin 2006, 181—205; eadem and Van Haeperen 2006. 



I INTRODUCTION 

was of course simply adopting terms current in late nineteenth-cen- 
tury Europe, together with the unself-conscious colonial baggage that 
attended them; in his own eyes, however, the point was to pick out and 
valorise the 'inepuisables reserves' of the ancient oriental civilisations 
and so reverse the conception, traditional since the Enlightenment, 
of a Roman Empire undermined and betrayed, above all in matters 
religious, by the Orient. Moreover, everyone now knows — could have 
known since the 1930s — that Cumont's oriental religions were to a 
very high degree hellenised and romanised, since otherwise they would 
never have been received. 3 They were thought of, and called them- 
selves, Egyptian, Phrygian and Persian, but the reality was far more 
complex. Nevertheless, in my view we need a word to denote a group 
of cults that were distinctive in the Graeco-Roman world, and which 
themselves emphasised their own alterity — even if we prefer to call it 
'peudo-alterity'. The conceptual loss that we incur by pretending on 
the one hand that these cults were just the same as any other religious 
association, the Poseidoniastai or the Mercuriales, or, on the other, that 
there is nothing to choose between Jewish or Christian groups and the 
oriental cults, because they are all 'oriental', seems to me to outweigh 
the trivial danger of being mistaken as a proponent of the conceptual 
follies of yester-year. 

Whatever the case with the ethnography of simple small-scale soci- 
eties, in the context of the Roman Empire, the term 'religion' must 
be used with care. It most clearly applies to what in Marxist terms 
we can call the entire, if historically changing, symbolic system(s) of 
the ideological superstructure. Admittedly, both 'Greek religion' and 
'Graeco-Roman religion' are elusive entities, which, for lack of control- 
ling ideas and dearth of information, no human being now can, or ever 
could, describe exhaustively, or even (one sometimes thinks) adequately; 4 



It must be said, however, that Cumont tried, albeit unsuccessfully, to distance himself 
from current prejudices: his very opening sentences allude to modern imperialism and 
European contempt for 'Levantines'. These expressions continued to have a long life: 
in 1928 we can find A.D. Nock writing: "The East conquered the West because it had 
something to give" (1962, 16). 

3 Cf. the review of Cumont 1929 by E.J. Bickermann, Orientalische Literaturzeitung 1931, 
210—13; also C. Schneider, Die griechischen Grundlagen der hellenistischen Religions- 
geschichte, ARW 36 (1939) 300-24. On this theme, see nowJ.-M. Pailler, Les religions 
orientales: troisieme epoque, Pallas 35 (1989) 95-113; Belayche 2000a, 2000b. 

1 The best effort is the two volumes of Nilsson GGR, first conceived in the 1920s and 
unlikely ever to be superseded; followed by Wilamowitz 1931-2. It is even legitimate to 
doubt whether a formula or set of formulae could be found that would lend coherence 



INTRODUCTION 6 

but no one seriously doubts that, in their Chinese-box manner, these 
symbolic systems provided a complex, differentiated, account of how 
the world functions and the rationale and goals of human existence. 
It is also pointless to object to the use of the word in relation to the 
individual city-states of the Archaic and Classical Greek worlds, or of 
Republican Rome, even though they lacked self-sufficiency as systems. 
But the worship of Mater Magna or Isis at Rome was only possible as a 
complement to other religious identities: these deities and the complexes 
of belief and praxis that accompanied them were excerpts or fragments 
from an alien religious world, a package in translation, offering a highly 
selective and instrumentalised glimpse of alterity. 5 No one could 'be' a 
Metroac or an Isiac in the same sense that one could be a Christian. 
At any rate, it seems evident to me that, with the exception of Judaism 
and the various Christianities, including 'gnosis', the religious institutions 
that entered the Graeco-Roman world from the eastern Mediterranean 
had no pretension, either organisationally or functionally, of offering 
a complete account of the world and the meaning of life, however 
they might have developed had Graeco-Roman paganism managed to 
endure as long as Hinduism. None of these movements can usefully be 
considered a self-sufficient or thorough-going religion in this sense. 'Cult' 
is therefore in many ways preferable. 6 As commonly used, cult denotes 
a specific form of worship subordinated to, and only making full sense 
within, a larger, composite structure, a religion. Applying such a view 
to antiquity, the word 'cult' would denote a lower level institution, in 
contrast to the larger totalities, Greek, Roman and Graeco-Roman, or 
'Hellenistic', religion. So far as it goes, this is acceptable: the oriental 



to the entire period from Mycenae to late antiquity. On the impasse of modern study 
of Greek religion, and esp. the ambivalent role of Nietzsche's Geburt der Tragodie (1872), 
see briefly A. Henrichs, Gotterdammerung und Gotterglanz: Griechischer Polytheismus 
seit 1872, in B. Seidensticker and M. Vohler (eds.), Urgeschichte der Moderne: Die Antike 
im 20. Jahrhundert (Stuttgart and Weimar 2001) 1-19. 

5 For this reason, I retain the traditional adjective 'oriental', on the grounds that, 
whatever the historical origins of these cults (especially that of Mithras), in a post- 
Saidian world it captures the appropriate ideological connotations of claimed alterity. 
I have deliberately refrained from decorating the adjective with scare-quotes. 

6 Reinhard Merkelbach, however, demonstratively entitled his first chapter on the 
Roman (as opposed to the Iranian and Commagenian) cult of Mithras "Die Mithras- 
mysterien — eine neue Religion" (1984, 75). Still more recently Robert Turcan, who 
used to speak of 'le mithriacisme' or 'les mysteres de Mithra' has recently chosen the 
expression 'la religion mithriaque' (2004: title). 



4 INTRODUCTION 

cults are best understood as options within a larger polytheistic system 
capable of virtually unlimited self-renewal. 

The main argument of the book is that the oriental cults, once 
adapted to a largely alien religious system in varying degrees recep- 
tive to them, became religious consumer-goods widely available for 
consumption within the host culture. In some ways they functioned 
like other group-cults in the Empire, providing an institutional structure 
for individual religiosity (cf. Riipke 2007b). Above all, they were not 
exclusive: they did not reject ritual modes proper to civic cult. They 
accepted the very essence of traditional religion, the bond established 
between sacrifice, as the central mode of ritual action, euergetic benefi- 
cence (in this case the distribution of food — sacrificial meat — supplied 
by the members of the elite and sub-elite who provided the victims), 
and the mass of the population, as a collective expression of common 
commitment to a highly asymmetrical political order. 

This however is only part of the picture. In the course of their 
Hellenisation, thanks to the far-reaching changes, or perhaps even 
genuine re-foundations, that they underwent, they developed their 
views of the cosmos in what we can only describe as an universalis- 
ing direction, which clearly distinguishes them from the group-cults 
of traditional civic deities. This development, partly at least at the 
level of myth, was intimately linked to their assimilation of the Greek 
institution best suited to the formation of an inner-directed religiosity, 
the mystery-cult, and backed by specific, and characteristic, demands 
upon the body and upon ethical character. These three features enjoyed 
a dialectical relation to that altered perception of the divine, broadly 
at odds with the Greek and Roman civic tradition, which the 'School 
of Leyden' has epitomised as 'servitude of God', and Paul Veyne as 
'fear of God', drawing inspiration from it but also strongly reinforcing 
it.' From a different perspective, J.Z. Smith has proposed the general 
term 'religions of "anywhere"' to denote a 'rich diversity of religious 
formations' in the Empire that occupied an interstitial space between 
domestic religion and civic religion. 8 All three of my cults, even that 
of Mater Magna, fall into this wider category, which also fits well with 



7 Cf. esp. Pleket 1981; Versnel 1981a; Veyne 1989 and 1999. 

8 J.Z. Smith, Here, There and Anywhere, in S. Noegel, J. Walker and B. Wheeler, 
Prayer, Magic and the Stars in the Ancient and Late Antique World (University Park PA 2003) 
21-36. 



INTRODUCTION D 

the problematic of the formation of an 'inner self in ancient religion. 9 
In relation to the perception of divine images, we might invoke what 
Jas Eisner represents as a new interiorised mode of visual experience, 
whereby the naturalism that is the visual correlate of civic religiosity 
recedes in favour of the construction of ritualised moments of 'contact- 
viewing', true encounters with god, of the kind repeatedly experienced 
by Lucius in Apuleius' Metamorphoses Bk. II. 10 This view is itself no 
doubt a version of J.-P. Vernant's conception of a long-term shift from 
mimesis to (pocvraoioc. 11 

In other words, the three oriental cults that concern me here were not 
simply absorbed into the larger system but showed themselves capable of 
quasi-autonomous development. 12 Indeed I would go so far as to claim 
that they played an important role in the transformation of religious 
sensibility, pin-pointed by the School of Leyden, that took place between 
the mid-Hellenistic period and the dominance of Christianity (say from 
the third century BC to the late fourth century AD), and more especially 
during the Principate. In Seneca's view, their adherents confused mere 
flamboyance with true religiosity, external trappings for true revelation 
(Logeay 2003, 33-6). We however are not compelled to follow him in 
this judgement, for he tacitly concedes that revelation was indeed their 
goal. I am very struck by the fact that it was the mysteries that most 
vigorously opposed the triumphant new cult and were therefore the 
target of the most bitter Christian diatribes (cf Turcan 1984; 1997). In 
a mere five centuries, these exotic, marginal cults, which had found it 



9 See J. Assmann and G.G. Stroumsa (eds.), Transformations of the Inner Self in Ancient 
Religions. Numen Book Series 83 (Leyden 1999); G.G. Stroumsa, From Master of 
Wisdom to Spiritual Master in Late Antiquity, in Brakke, Satlow and Weitzman 2005, 
183-96; AA.VV, Pagani e cristiani alia ricerca della salvezza (secoli I-III). XXXIV incontro 
di studiosi dell'antichita cristiana, Roma, 5-7 maggio 2005 (Rome 2006). 

10 J. Eisner, Cult and sculpture: sacrifice in the Ara Pacis Augustae, JRS 81 (1991) 
49-61; idem 1995; cf. his collected and new essays, Roman Eyes: Visuality and Subjectivity 
in Art and Text (Princeton 2007), esp. the epilogue: From Diana via Venus to Isis; some 
reservations in R. Brilliant, Forwards and Backwards in the Historiography of Roman 
Art, JRA 20 (2007) 7—24 at 17. My account can perhaps be seen as a partial comple- 
ment to Eisner's at the level of specialist religious institutions. Note also D. Freedberg, 
The Power of Images: Studies in the History and Theory of Response (Chicago 1989). 

11 J.-E Vernant, De la presentification de l'invisible a l'imitation de l'apparence, in 
idem, Entre mythe et politique (Paris 1996) 359—77 at 376f. (first published 1983). 

12 To that extent, I may be thought to be less indebted to Cumont, whose notion of 
mystery was always quite vague, if not muddled, than to Richard Reitzenstein's notion 
of 'Hellenistische Mysterienreligionen' (Reitzenstein 1927). Unfortunately, the more 
'oriental' texts he found, the more his concept sank into over-detailed incoherence. 



b INTRODUCTION 

difficult to gain admittance to the Roman world, became the paladins 
of the defence of the traditional paganism of the state. 

To put the matter in very general terms: the ancient normative reli- 
gious structure, what E.R. Dodds once called (in relation to Greece) 
the 'inherited conglomerate', altered significantly under the pressures 
brought to bear on the ideological realm by the socio-political order. Of 
primary importance here was the fact that, as a result of the Roman 
Republican elite's incessant conquest and incorporation of new terri- 
tories, and the subjection of their inhabitants to Roman exploitation, 
the society of the Empire was very different from that of Republican 
Rome. This development, which was of course preceded and indeed 
partly enabled by Alexander's conquests and the expansion of the Hel- 
lenistic Greek world in the form of tax-based complex states, provided 
the context for the gradual culture-contact that brought about those 
changes in provincial culture and society we call, for want of a better 
term, Romanisation but also, conversely, profoundly affected Roman 
culture too. 13 From this point of view, I would claim, the oriental cults 
were a destabilising factor within the traditional religious structures of 
the Graeco-Roman world, for all that these had been re-activated in a 
conservative sense, by political fiat, during the Augustan period. 14 Their 
success was due to a variety of factors, but I see a common denomina- 
tor in their offer of institutional, and partly routinised, access to the 
subjectively appealing role of servant of God through the medium of 
secret revelation, the magna religionis silentia of Apuleius, Met. 11.21. 

I am of course aware that this programme, my use of the term 
oriental cults, and no doubt my attempt to save at least aspects of the 
category of the dying-rising god, introduced by Frazer and re-worked 
for the Graeco-Roman period by Cumont, must appear, especially in 
the Anglo-American world, in many ways reactionary 15 It is there that 



13 The most important recent attempt to come to terms with these changes in the 
religious context is the project directed from Tubingen by H. Cancik, which includes 
22 sub-projects and which came formally to an end in 2007: cf Petzold, Riipke and 
Steimle 2001. The most important conference proceedings to emerge from this proj- 
ect so far are Cancik and Rupke 1997; Cancik and Hitzl 2003; Cancik, Schafer and 
Spickermann 2006. But there is also much of value in the relevant chapters of Beard, 
North and Price 1998. 

14 Cf. A. Wallace-Hadrill, Rome's cultural revolution, JRS 79 (1989) 157-64. 

15 This category seemed to have been totally discredited by the famous article 
of J.Z. Smith 1984. I have taken heart from two books that appeared in 2001: the 
scrupulous survey of Mettinger 2001, who, without wanting to revive Frazer, provides 
a substantial defence of the seasonal interpretation of dying and rising gods in the 



INTRODUCTION 



the influence of Ramsay MacMullen's Paganism in the Roman Empire, with 
its direct attack upon the Cumontian grand narrative and its stress on 
the centrality of civic religion in the Latin-speaking west, has been 
most deeply felt. It was followed shortly afterwards by the publication 
of Robin Lane Fox's similar demonstration, heavily influenced by the 
epigraphic work of Louis Robert, of the continuing vitality of civic 
cult in the eastern Mediterranean up to and well beyond the age of 
Constantine. 16 Neither book was at first much noticed on the European 
continent, 1 ' where the influence of the various branches of the History 
of Religions school, especially the School of Rome (notably Raffaele 
Pettazzoni, Ugo Bianchi), and of Maarten J. Vermaseren's enormous 
series, Etudes preliminaries aux religions orientales dans l'Empire 
romain, remained dominant and continued largely to underwrite the 
traditional grand narrative. In the very year that MacMullen's Paganism 
appeared, Vermaseren edited an important collective volume intended 
to up-date Cumont's notion of the oriental religions by including Juda- 
ism and Christianity as well as the traditional components: Cybele, Isis, 
Mithras, the Syrian cults, magic and astrology (Vermaseren 1981b). 18 



ancient Near-East: Ugaritic Baal, Melqart-Heracles, Adon(is), Eshmun-Asclepius, but 
also Osiris and Dumuzi-Tammuz; and some of the essays in Xella 2001 (on Osiris, 
Adonis, Attis etc.). For acute appreciations, see C. Bonnet, rev. of Mettinger, Biblica 84 
(2003) 587—91; eadem, La mort des hommes, la mort des dieux. Reflexions autour de 
quelques livres recents, LEG 71 (2003) 271-80; of Xella's collection: T.N.D. Mettinger, 
History of Religions 43 (2003-4) 341-3. 

16 MacMullen 1981; Lane Fox 1986; cf. too Trombley 1993-4. 

'' MacMullen 1981 was translated into French as Paganisme dans l'Empire romain in 
the series Chemins de l'histoire (PUF) in 1987; MacMullen 1984 into Italian in 1989; 
MacMullen 1997 into French in 2004. Lane Fox 1986 was translated into French as 
Paiens et chretiens (Toulouse 1997). So far as I know, these are the sole translations into 
Continental European languages. G. Alfbldy argued much the same in 1989, e.g. "Statt 
von einem 'Sieg des Orients' konnte man beinahe von einem 'Sieg des Okzidents' in 
der mittleren Kaiserzeit sprechen" (1989, 78). 

18 Esp. the contributions by R. Van den Broek, Friihchristliche Religion, pp. 363—87; 
Kotting 1981; G. Quispel, Gnosis, pp. 413—35; A. Boehlig, Manichaismus, pp. 436—58; 
H. Kiinzl, Judentum, pp. 459—84; and M. Oppermann, Thrakische und Danubische 
Reitergotter, pp. 510-36 (Cumont was interested in these last, but the crucial, and rather 
poor, publication by D. Tudor appeared only in 1937: see his letter to RostovtzefF of 
Nov. 9, 1937 = Bongard-Levine 2007, 223-5 no. 127 with fn. 860). Cumont had himself 
widened his category in the fourth ed. of 1929 by including the Dionysiac mysteries, on 
the specious grounds that they were 'demi-orientalises'. For a sympathetic appreciation 
of Cumont's achievement, see R. Turcan, Franz Cumont, un fondateur, Kernos 1 1 (1998) 
235—44; on his wide influence, note B. Rochette, Pour en revenir a Cumont. . .L'oeuvre 
scientifique de Franz Cumont cinquante ans apres, in Bonnet and Motte 1999, 59—80; 
and the bibliometric study of G. von Hooydonk and G Milis-Proost, The Scientific 
Survival of Franz Cumont, in Bonnet and Motte 1999, 81—91. 



8 INTRODUCTION 

Moreover the public appetite for the grand narrative has persisted: 
one of the most successful books of recent years in this area, closely 
modelled on Cumont's original project, has been Robert Turcan's Les 
cultes orientaux dans ['Empire romain, which has seen three French editions 
and, under the rather odd title The Cults of the Roman Empire, been widely 
sold in English translation as a student text (Turcan 1992a; 1996a). 19 
The old grand narrative is still alive and well in more traditional New 
Testament circles. 

Nevertheless, the scholarly retreat from Cumont's idea of 'oriental 
religions' is clear. It has seemed safer, less contentious, to work on just 
one or other of the cults concerned rather than attempt a synthesis. 20 
The direct effect of MacMullen's book was to minimise the historical 
importance of these cults vis-a-vis civic cult, and thus undermine the 
validity of the traditional assumption of a link between their success and 
that of Christianity. Explanations of the latter had to take a different 
form. 21 This partly explains the recent interest, especially in England, 
in the rise of monotheism. 22 Open dissatisfaction with the Cumontian 
grand narrative has now spread to the European continent, as the series 
of recent conferences in connection with the centenary of the first 
edition of Les religions orientates (1906) has made plain. 23 Perhaps the 



19 Cf. Liebeschuetz 1992, 250—62. Note also several books issued by reputable 
publishers and widely circulated: Giebel 1990, which runs through Eleusis, Dionysus, 
Samothrace, Cybele, Isis and Mithras (in that order); H. Kloft, Mysterienkulte der Antike: 
Gutter, Menschen, Rituale 3 (Munich 1999); R. Girault, Les religions orientates (Paris 1995); 
A. Tripolitis, Religions of the Hellenistic-Roman Age (Grand Rapids MI and Cambridge 
2002); P. Scarpi (ed.) Le religioni dei misteri (Milan 2002). 

20 E.g. on Isis: Dunand 1973; Merkelbach 1995 (highly idosyncratic); Takacs 1995 
RICIS; on Mater Magna: Vermaseren 1977; CCCA; Roller 1999; Borgeaud 1996/2004 
on Mithras: Merkelbach 1984 (again highly idosyncratic); Clauss 2000; Turcan 2000 
Ulansey 1989 and Beck 2006 are exclusively concerned with astral interpretation. 

21 E.g. MacMullen 1984; 1997; idem, Constantine (New York 1969); Lane Fox 1986; 
R. Stark, The Rise of Christianity: How the obscure, marginal Jesus Movement became the Dominant 
Religious Force in the Western world in a few Centuries (Princeton 1996); Hopkins 1999. 

22 E.g. Athanassiadi and Frede 1999. S. Mitchell runs a post-graduate project on 
this topic at the University of Exeter, UK. 

23 Bonnet and Bendlin 2006; Bonnet, Riipke and Scarpi 2006; Bonnet, Ribichini 
and Steuernagel 2008; Bonnet, Pirenne-Delforge and Praet 2009. Some of the papers 
contributed were however not published in the various Proceedings. The process of 
revaluation was initiated by the publication of much of Cumont's academic correspon- 
dence (Bonnet 1997; 2006; also now Bongard-Levine et al. 2007) and the conference 
on syncretism at Rome in 1997, organised by C. Bonnet and A. Motte in honour of 
the fiftieth anniversary of Cumont's death (Bonnet and Motte 1999); note also the 
helpful sketch of Cumont's intellectual contacts, esp. with German scholars such as 



INTRODUCTION y 

most important achievement of these meetings has been to expose the 
ideological underpinning, from Friedrich Creuzer to Droysen, Momm- 
sen and Cumont, of the very notion of an 'oriental cult'. 24 Although 
some older scholars, such as Marie-Francoise Baslez, retain the tradi- 
tional category, it is noticeable that virtually no one addresses the issue 
of their specific appeal: intellectual energy has been directed rather 
to reconsidering the strengths and weaknesses of the notion of civic 
religion, the problems involved in the notions of interpretatio Graeca and 
Romana, and, most strikingly, the expansion of the term to include not 
merely the Dionysiac mysteries but also the Jewish diaspora and early 
Christianity. The category, already capacious enough, now threatens 
to lose such coherence as it could ever claim to. In studies of local 
centres, the oriental gods disappear among a plethora of other deities. 25 
Moreover, it is now becoming usual to treat the temples and shrines of 
these cults as elements of a general Graeco-Roman paganism rather 
than as a distinctive architecture. 26 Some at least of the objections, 
explicit and implicit, to the grand narrative are sound, since the intel- 
lectual claims of these cults were indeed rather different from what late 
nineteenth-century scholars believed. There is also a legitimate debate 
to be conducted over the appropriateness of Cumont's wide — in his 
own day, quite traditional — demarcation of the 'oriental religions'. I 
do not however wish to embark on this task in the present book — the 
theme has after all been widely aired in the past few years. 



T. Mommsen, Usener, Diels and Hirschfeld, in Bonnet 2006. An important predeces- 
sor however is A. Rousselle, La transmission decalee: nouveaux objets ou nouveaux 
concepts, Annates ESC 1989, 161-71. 

24 Esp. P. Borgeaud, L'Orient des religions. Reflexion sur la construction d'une 
polarite, de Creuzer a Bachofen, in Bonnet and Bendlin 2006, 153—62; E Payen, Les 
"religions orientales" au laboratoire de l'Hellenisme, 1: G. Droysen, in Bonnet and 
Bendlin 2006, 163—80; also the very valuable essay by Bonnet and van Haeperin 
2006. The role of PEgypte imaginaire, though often neglected, is of great importance 
here, cf. E. Hornung, Das esoterische Agypten (Munich 1999); W. Seipel (ed.), Agyptomanie: 
Europaische Agyptenimagination von der Antike bis Heute. Schriften des Kunsthistorischen 
Museums, Wien 3 (Vienna 2000). 

25 E.g. Spickermann 2003; idem, Mogontiacum als Zentralort der Germania Superior, 
in Cancik, Schafer and Spickermann 2006, 167—93; C. Tsochos, Philippi als stadtisches 
Zentrum Ostmakedoniens in der hohen Kaiserzeit: Aspekte der Sakraltopographie, 
ibid. 245-72. 

26 E.g. Egelhaaf-Gaiser 2000; eadem, Roman Cult-Sites: A Pragmatic Approach, 
in Riipke 2007d, 205-21; eadem and Schafer 2002; Steuernagel 2001; 2004; 2006; 
M. Trumper, Negotiating Religious and Ethnic Identity: The Case of Clubhouses in 
late Hellenistic Delos, in Nielsen 2006, 1 13-40. 



1 INTRODUCTION 

When I use the term oriental cults in this book I mean not Cumont's 
broad notion, itself taken over from the nineteenth-century historiog- 
raphy on the topic, but simply the three major movements that seem 
to me to possess a typological coherence. 2 ' The wider category he used 
thus has no special significance for me and I make no attempt here to 
defend it. The issues are complicated by the nature of the evidence 
and of the subject itself: there is no easy way of resolving them that 
does not do violence to that complexity. My focus, as I make clear in 
Chapter 1, is upon those cults that, originating in the eastern Medi- 
terranean, underwent a thorough-going transformation in the process 
of adapting themselves to Hellenistic-Roman culture, in particular by 
assimilating features characteristic of the grand mysteries celebrated 
from time immemorial at Eleusis. The most important of these were 
initiation, as a ritual of dis-aggregation from the crowd, and the vow 
of silence regarding what had been communicated in that experience. 28 
The originally Phrygian cult of Cybele and Attis, and the Egyptian cult 
of Isis, Osiris and Serapis in particular conform to this model. 'Persian' 
Mithras however did not. 29 One of the questions I ask is why it was 
that the initiatory experience, particularly in its Eleusinian form, gained 
such a hold at this period. There certainly were other Graeco-Roman 
mysteries, such as the Samothracian mysteries, those of Dionysus, of the 
Cabiri at Thebes, of the Roman emperor, of numerous local divinities 
such as Zeus of Panamara, but neither their content nor what we know 
of their ritual praxis seem to have had any influence on my group of 
three. This justifies my decision not to include the Dionysiac myster- 
ies, which were not perceived as alien and did not need to assimilate 
themselves to a foreign culture. 30 They form part of the socio-religious 



27 I here rely upon the framework established by Chirassi Colombo 1982. She 
employs the periphrasis "i tre piu noti sistemi misterici dei primi secoli dell'impero", 
for which my short-hand is "the oriental cults". 

28 There have been a number of good recent accounts of aspects of Graeco-Roman 
initiation: Meslin 1986; Turcan 1996c; Riedweg 1998; and the collection edited by 
Moreau 1992a,b. These have largely superseded older accounts from the History of 
Religions School, such as Bleecker 1965. 

29 Cf. the interesting recent description by Turcan: "(La religion mithriaque) ou con- 
fluent de fait un lointain heritage iranien, une cosmologie fortement impregnee d'idees 
grecques (stoiciennes en particulier) et la culture romaine ambiante" (2004, 259). 

30 On Greek mysteries, see recently, apart from Burkert 1987, the essays in Cosmo- 
poulos 2003; fundamental on Samothrace: Cole 1983. Several scholars, such as M. Jost, 
Mystery Cults in Arcadia, in Cosmopoulos 2003, 143—68, and A. Schachter, Evolutions 
of a Mystery Cult: The Theban Kabiroi, in Cosmopoulos 2003, 1 12-42, have stressed 
the fact that in the Greek world such mysteries were generally closely connected to 



INTRODUCTION 1 1 

context within which the oriental cults flourished rather than part of 
my subject here. 

As I try to make clear in Chapter 4.2, I do not think the relative 
success of the oriental cults in the Empire was a sign of dissatisfaction 
with the religious options on offer, whether at the level of public-civic 
cult, institutionalised group worship, or household observance. It seems 
clear that 'traditional' paganism was sufficiently diverse and complex 
to offer something to almost everyone. The attraction of the oriental 
cults, and their mystery component in particular, cannot be as easily 
explained as Reville, Duruy Cumont and the others imagined. Among 
these factors we can surely list the growth of a 'consumer' demand for 
more differentiated religious experience between the household and 
public cult; the gradual establishment of the idea that religious experi- 
ence was of a special order, not reducible to other social experiences 
and roles; and the demand in a dynamic and complex society for new 
opportunities for symbolic self-expression and the exchange of new 
types of specialised knowledge. Then again, there are signs of larger 
shifts in this period relating to the value and place of the body, and 
the negotiation of the idea of selfhood, in the religious context, which 
are most easily grasped through the metaphor of slavery to a god. If 
I sometimes summarise these themes by speaking of anxieties in the 
face of the socio-cultural changes brought about the Empire, I do so 
not to hark back to the psychologism of a Dodds, but to express my 
conviction that the oriental cults were simultaneously a sign of their 
times and active participants in the creation of what we might call 
religious 'density' in the Empire. 

There is in fact another reason for my choice, which I may perhaps 
state here. This is that many of the ancient sources refer to the cults 
of these deities by the general term 'mysteries'. Non-Christian authors 
of the mid-Principate as well as the Christian apologists agree on this 
point. The insistence on differences and distinctions between them 



civic cult (e.g. the third-century BC rebuilding of the Kabeirion was financed by the 
city of Thebes), or at any rate carried out in the context of the civic calendar, as was 
clearly the case at Panamara: P. Roussel, Les mysteres de Panamara, BCH 27 (1927) 
57-137 (here the word 'mystery' simply meant a public feast under the aegis of Zeus). 
For a recent account of Dionysiac initiation in the later period, see Turcan 1992b and 
2003; H. Schwarzer, Die Bukoloi in Pergamon. Ein dionysischer Kultverein im Spiegel 
der archaologischen und epigraphischen Zeugnissen, in Nielsen 2006, 153—68. On the 
theories of G. Sauron and P. Veyne relating to the Villa of the Mysteries, cf. J.-M. 
Pailler, Mysteres dissipes ou mysteres devoiles, Topoi 10.1 (2000) 373-90. 



1 2 INTRODUCTION 

makes it difficult to acknowledge the extent to which the goods on offer 
in the ideological and religious market-place of the Empire resembled 
one another. 31 These goods aimed at satisfying what I still think of as 
new spiritual needs that they themselves encouraged and fostered, and 
by which they were sustained in return. Christianity shared with the 
mysteries, if nothing else, a specific geographical and cultural context. 
To my mind, the reluctance of many scholars to study the whole picture 
does not foster our understanding of the grander historical processes at 
work. 32 We must be allowed to try to discover whether studying a group 
of cults that bear an obvious family resemblance to one another, and 
have not been chosen arbitrarily, helps in the task of understanding the 
shifts in religious thought of a period in which the political order was 
desperately on the look out for effective ideological allies. 

From this it will be clear that the category 'mystery-religion' plays a 
central role in my conception of the oriental cults. 33 What is most truly 
characteristic of the latter is not so much their public face, though I 
pay considerable attention to this, as their self-representation as mys- 
teries. This term is used by ancient writers to refer to a complex of 
knowledge, understanding and action that was restricted to initiates, and 
might not be divulged to non-initiates. The rule of silence applied to 
what the teacher, the mystagogue, revealed to the mystes, the neophyte. 
Initiation was thus an effective means of social control: our ignorance 
of the teachings of these cults is due to the fidelity of their adherents 
to the rule. 34 Whatever the details, however, the core idea was that 
submission to the power of the god was met by the individual offer of 
well-being in this world and, from the first century AD, salvation in 
the world beyond. The idea of a central revelation or secret, and of a 



31 For the idea of a religious market-place, which I find attractive, see North, 1992; 
idem, 2003. 

32 See my paper given at the conference in Rome (Nov. 2006): "Le pouvoir des 
concepts et la valeur de la taxonomie: religions orientales, cultes a mysteres", to appear 
in Bonnet, Pirenne-Delforge and Praet 2009. 

33 I here tend to follow U. Bianchi, The Religio-Historical Question of the Myster- 
ies of Mithra, in Bianchi 1979, 3—68; also Bianchi and Vermaseren 1982. Although 
she is the most eminent pupil of U. Bianchi, Sfameni Gasparro has argued that the 
evidence for mysteries in the cult of Cybele and Attis is doubtful and anyway requires 
strict scrutiny as regards the claims and offers made (Sfameni Gasparro 1985). I would 
resist this conclusion. 

31 Eusebius of Caesarea, at the beginning of the fourth century AD, alludes to 
this rule in the Orphic text from Aristobulus that I cite in the epigraph on p. v. As a 
Christian, he was not bound by the prescription, but the fear of such sacrilege had 
penetrated deeply into the collective mentality. 



INTRODUCTION 1 3 

process of initiation to prepare for its apprehension, are the fundamental 
concepts that define the mysteries. I see this notion of a central secret 
however mainly as a form of inducement to, and reinforcement of, a 
specific set of ethical commitments that set these cults apart from the 
practice of what, for want of a better term, we may call routinised, 
common-or-garden paganism. 

The oriental cults, particularly the mysteries they offered, have from 
antiquity widely been regarded as a sort of preparation for the success 
of Christianity. As regards its origins and typology, Christianity was an 
oriental religion, and for their part the apologists were astonished — hor- 
rified — to discover the degree to which the 'true religion' resembled the 
mysteries, a fact that they could only explain by invoking the idea that 
the Devil had invented the similarities in order to discredit the 'Christian 
mystery'. Historically speaking, however, we have to do with a situation 
of competition between different options, closely related in that they 
stemmed from analogous cultural systems in the eastern Mediterranean 
and were subject to the same pressures. Christianity and the mysteries 
resemble one another because they offered analogous solutions to the 
needs generated in certain sectors of the society of the Roman Empire. 
"The Epistle of Aristeas, the Acts of the Apostles, and the Golden Ass 
by Apuleius, here especially the eleventh book, are only a few of many 
documents to demonstrate missionary interest in the well-being of all, 
where the propagated religion is proposed as the major instrument for 
bringing this about and for serving the best interests of the individual 
as well." 33 These new dissatisfactions and desires were the result of the 
Romans' assimilation of territories into their own cultural system, and 
the gradual homogenisation of cultures, despite the continuing differ- 
ences that served to distinguish the innumerable subject communities 
of the Empire. 36 

In view of all this, the best way of thinking about religious change 
in the Roman Empire is not to understand Christianity as having suc- 
ceeded thanks to the praeparatio evangelica, the preparatory work effected 
by the mysteries, but as one option among many others with similar 



35 D. Georgi, Socio-economic Reasons for the 'Divine Man' as a Propagandistic Pat- 
tern, in Schussler Fiorenza 1976, 27—42 at 36. As is well known, M. Goodman, Jewish 
Proselyting in the First Century, in Lieu, North and Rajak 1992, 53—78, and elsewhere, 
has argued strongly against the claim that the Jews actively proselytised. 

!li The complexity and ambiguity of the process of 'Romanisation' (a term he rejects) 
has been recently underscored by Hingley 2005. 



14 INTRODUCTION 

features that came into being at a particular historical moment. 37 The 
issue of how Christianity established itself in this context, which of 
course included the existence of the oriental cults, is a different one, and 
requires its own separate analysis. I agree with the scholarly consensus 
on this. In any case, the main role of the mystery-component of the 
oriental cults lay in the general belief that they had contributed to the 
triumph of Christ rather than in any intrinsic significance they may 
have had in this regard. 

The purpose of this brief introduction has been simply to set out 
the larger conceptual framework of the book, whose organisation 
corresponds pretty exactly to the ideas set out here. The short first 
chapter sets out the three 'orders of things', cosmic, socio-productive, 
and eschatological, that taken together establish the sub-system of 
belief within the wider framework of a religious institution. Chapter 2 
presents the three cults' systems of beliefs at some length in terms of 
these three orders. Here I attempt, in defiance of our general igno- 
rance, to piece together the shards of their teaching and set them into 
a coherent system of explanation. Chapter 3 discusses their systems of 
values, which I see as playing a major role in their appeal. The longest 
section of the book, Chapter 4, is devoted to an examination of the 
public, exoteric, rituals in which the adherents of the two major cults, 
Mater Magna and Isis, displayed both their self-proclaimed exoticism 
and their assimilation of the values of the civic community. Beyond 
that, I do what I can with the evidence concerning initiation. Despite 
the successful insistence on secrecy, I believe we do have enough scraps 
of information to give us a rough idea of what such rites consisted in, 
though of course without the details. It is hoped that the reader will 
be able to form his or her own impression of the case, and come to his 
or her own conclusion about the religious character of the three cults 
and the legitimacy of my decision to take them together. 38 The final 



37 Subjectively of course the difference was, or might be made to appear, total. The 
rapidity with which early Christianity defined itself against paganism in terms of the 
issue of sacrificial meat is made clear by Woyke 2007. 

38 I accept of course that this is procedure is not self-evident; but think there still is 
a case to be made in favour of the procedure advocated long ago by H. Willoughby: 
"The most nearly exact procedure would seem to be to emphasize those fundamental 
aspects of the mystery type of religion which were characteristic of all the cults in 
common and to balance this with a detailed investigation of the idiosyncrasies of each 
particular cult": Pagan Regeneration (Chicago 1929) 33, cited by Wiens 1980, 1249. 



INTRODUCTION 1 5 

chapter is an inevitable outcome of what the book sets out to achieve, 
in that it provides the tools needed to make a considered judgement 
of the relation between the oriental cults and Christianity. 

I often refer in this book to social conflicts, the difficulties the Roman 
Empire faced in trying to integrate the system it had created, the con- 
tradictions both personal and collective generated within it, and the 
role played by various institutions, particularly religious ones, in allevi- 
ating these problems. The epigraph to this Introduction is taken from 
Ernest Renan (1823-92), whose Vie de Jesus first appeared in 1863 and 
within a year had been reprinted thirteen times. Despite the overstate- 
ment, Renan's claim must cause us to reflect on the subjectivity of any 
historical engagement with the past. No one can unthink the cultural 
paradigms within which he or she has been socialised. It was with this in 
mind that I have chosen a 'parabolic' approach: the citation from Renan 
serves to illustrate my conviction that historical analysis is primarily an 
individual or personal matter. I would even go so far as to claim that 
the historian's objectivity is a mere fiction serving to mask the author's 
constant and undeniable presence, both willed and unwilled, conscious 
and unconscious, in his or her text. 39 I hope that in the course of this 
attempt to get at the truth, I will be able to persuade the reader of the 
legitimacy of my particular way of looking at things. 

Habent sua fata libelli. Indeed; but my ideal reader would have taken 
time beforehand to re-read the chapters of Apuleius, Metamorphoses Bk. 
11.19-30 that constitute what is by far the most valuable surviving 
attempt to re-create the nervous apprehension, the febrile intensity, 
the self-abasement and sheer gratitude that ideally accompanied the 
experience of encountering the magna religionis silentia evoked in the sec- 
ond epigraph on p. v, the grand, ineffable revelations of the mysteries. 
Whatever can be said in favour of an ironic reading of Bk. 1 1 , and 
there is no doubt plenty, it seems to me that, if we are to do justice to 
the oriental cults, we would do well to make the effort of learning to 
value Mithras' words as a fictional, and of course idealised, insight into 
a particular, paradigmatic, frame of mind. This book is a circuitous, 



39 Cf. the remarks of Hingley 2005, 1—13, and the interesting work on the twentieth- 
century constructions of ancient slavery by N. McKeown, The Invention of Ancient Slavery? 
(London 2007). 



1 6 INTRODUCTION 

academic attempt to persuade the reader to do just that. Criticism of 
the old grand narrative, partly justified though it be, is in danger, I 
think, of ignoring or misrepresenting the fascination exercised by the 
mysteries' offer of a privileged relation to divinity culminating in a 
privileged fate after death. 



CHAPTER ONE 
RELIGION, CULT AND MYSTERY 



a>q xa . . . &7toppr|Ta xr\q Kara ra iruatfipia xekexr\q 
ev8o[^ Jotepov xe Koci ae|iv6tepov . . . tow 6eoiv 
a7io8o0oir| . . . 

Letter of Commodus to the Eumolpidae at Eleusis, 
AD 180-92: Syll. 873 11. 5-13 

Le sacre revient au galop . . . Heureux, parce qu'il 
est temps de sortir de la mise a plat systematique. 
Dangereux, parce que rien n'est plus nocif que le 
faux sacre. 1 

Haben jene Nationen der Vorwelt einander nur 
Elephantenzahne zugefuhrt, und Gold und 
Sklaven? Nicht auch Erkenntnisse, religiose 
Gebrauche und Gotter? 

E Creuzer, Symbolik und Mythologie der alten Volker, 
besonders der Griechen (18 10-12') Introd. 

In the Introduction I have already set out my reasons for choosing 
the group of cults that are the subject of this book. We now must try 
to enter their conceptual world, comprehend their rituals. The very 
fact that they managed to expand all over the Roman world confirms, 



1 J. de Bourbon-Busset, Localiser le sacre, in Champs du sacre (Paris 1982) 3. I have 
hesitated over the appropriateness of this epigraph, but its oddity finally induced me 
to include it. My objection is not so much that everyone writes whatever comes into 
his head, as that supposedly specialist publications appeal to esprits forts to discuss the 
idea of the sacred on the basis of supposedly objective knowledge (e.g. J. Ries, Retour 
ou permanence du sacre? in idem 1986a, 1—13). To be specific: what does it mean 
systematically to flatten everything, in connection with the sacred? Is it certain that 
there is 'nothing more dangerous' than the pseudo-sacred? And who exacdy has the 
right to decide what the 'pseudo-sacred' might be? My reservations are by no means 
rhetorical; as good a historian of religion as C.J. Bleeker once wrote an article entitled, 
How to distinguish between True Religion and False Religion (in idem, 1975, 67—75), 
whose last paragraph reads: "One knows the tree by its fruits. The good deeds prove 
the truth of religion . . . we are not totally deprived of criteria to distinguish true religion 
from false religion. But such criteria should be handled with wisdom and discretion". 
If he had known Karlheinz Deschner's vast Kriminalgeschichte des Christentums (8 vols, up 
to 2004) (Reinbek/Hamburg 1986-), or M.J. Engh, In the Name of Heaven: 3,000 Years 
of Religious Persecution (Amherst 2007) 89-140, 161-213, 226-250, he would probably 
not have expressed himself so blithely. 



1 8 CHAPTER ONE 

whatever the actual number of their adherents at any one time, that 
the populations of quite diverse areas of the Empire felt attracted to 
such religious experiences and found in them a means, alongside more 
traditional religious forms, of responding to some of the new anxieties 
lurking in the realm of the imaginaire. 

The task of studying a heterogeneous group of cults like this is stud- 
ded with difficulties. In order to provide a self- consistent and coherent 
account, I have adopted a method I call 'rapprochement', which under- 
lies the book's structure. This procedure requires a little explanation. 

In my view, religion is a cultural system itself articulated in a variety 
of sub-systems that reflect, at the level of the imaginaire, the real con- 
ditions of existence in a specific historical formation. 2 Every society, 



2 See C. Geertz, Religion as a Cultural System, in: M. Banton (ed.), Anthropological 
Approaches to the Study of Religion (London 1966) 1-46, as an example of such a substan- 
tive definition. M. Spiro, Religion: Problems of Definition and Explanation, in: ibid. 
85—126, offers a functional one. Geertz is, with some justification, the most influential 
older anthropologist of religion, though has more recently been subjected to criticism. 
Some sympathy for Durkheimian sociology too is surely indispensable: its rejection 
of psychological accounts, for example, the conception of religion as the administra- 
tion of the sacred, or the analysis of religious phenomena as founded on a mixture 
of myth and ritual (and its analogies): De la definition des phenomenes religieux, 
LAnnee Sociologique 2 (1899) 1-28, with S. Lukes, Emile Durkheim (Harmondsworth 1973) 
237—44, 450-84. Phenomenological views, as represented by G. van der Leeuw, Reli- 
gion in Essence and Manifestation: A Study in Phenomenology, tr. J.E. Turner (London 1938, 
1964 2 ) or C.J. Bleeker, The Sacred Bridge. Researches into the Nature and Structure of Religion 
(Leyden 1963), are less productive. A preliminary account of the problems involved in 
defining religion is offered e.g. by M.B. McGuire, Religion. The Social Context (Belmont, 
CA 2002 5 ) 5—12; more rewardingly, Y. Lambert, La 'Tour de Babel' des definitions 
de religion, Social Compass 38.1 (1991) 73-85, or — in more detail — J.V Spickard, A 
Revised Functionalism in the Sociology of Religion: Mary Douglas' Recent Work, 
Religion 21 (1991) 141—64. Also worth mentioning are: R.B. Finnestad, Religion as a 
Cultural Phenomenon, in: G. Englund (ed.), The Religion of the Ancient Egyptians. Cogni- 
tive Structures and Popular Expressions (Uppsala 1987) 73-96; and the global reflections of 
G. Flood, Beyond Phenomenology: Rethinking the Study of Religion (London and New York, 
1999). Of great value for assessing the present state of play are two volumes published 
as Supplements to Numen: TA. Idinopoulos and B.C. Wilson (eds.), What is Religion? 
Origins, Definition, and Explanations. Studies in the History of Religions (Numen Book Series, 
81) (Leyden 1998), andJ.G. Platvoet and A.L. Molendijk (eds.), The Pragmatics of Defining 
Religion. Contexts, Concepts and Contests. Studies in the History of Religions (Numen Book 
Series, 84) (Leyden 1999). Among readings in rough conformity with mine, see B.S. 
Turner, Religion and Social Theory. A Materialist Perspective (Guildford 1983, 1991 2 ), and 
Gordon 1979a. He there defines the interdependence of religion and other cultural 
systems: "For the cognitivist, religion is one of the central and most important codes 
in a society. That is not, of course, to say that it is in any way autonomous, since it 
is linked in all sorts of ways to others, alimentary, reproductive, linguistic codes, for 
example. But it is central (especially in a traditional society) insasmuch as it articulates 
for its adherents the scope and nature of power, both 'vertically' between us and the 



RELIGION, CULT AND MYSTERY 1 9 

therefore, produces a cultural construction whose ambition it is to 
describe and taxonomise perceived reality, not merely at the level of 
empirical facts but also at the level of the imaginaire. During the process 
of acculturation from childhood, this heritage is transmitted to all the 
members of the community who absorb it, and are thus enabled to 
sidestep the burdensome intellectual challenge of developing a personal 
or private explanation of the order of the real and its counterpart in the 
imaginaire. Myths constitute the vehicle of transmission of this collective 
account, which undergoes adjustment corresponding to changes in the 
objective conditions of the historico-social formation in which they are 
elaborated. It thus becomes a cultural mediator of great importance, 
offering a shifting form of explanation that allows the members of 
the community to face, without excessive anxiety, a reality that might 
otherwise appear chaotic and uncontrollable. 3 



divine world, and 'horizontally' between men" (p. 17); "We can perhaps understand 
religion, in this context, as a specialized sub-system of language focused upon the 
category 'power'" (p. 19). 

! Here I can name some of my debts; whether or not I agree with what they say, all 
have contributed in one way or another to forming my position. I would note especially 
the series of suggestive introductions by G. Filoramo to the excellent re -publication of 
R. Pettazzoni's old 6-vol. collection Miti e leggende (1948-63) with the same title (Turin, 
undated but ca. 1990-91); E. Durkheim, The Elementary Forms of the Religious Life (1912, 
tr. E.W. Swain, London 1915), Book II, chap. 3; G.S. Kirk, Myth, its Meanings and Func- 
tions in Ancient and other Cultures (Cambridge 1970); and A.E.Jensen, Myth and Cult among 
Primitive Peoples (Chicago 1973). The relations of mutual dependence between myth and 
ritual have been examined critically by J. Fontenrose, The Ritual Theory of Myth (Berkeley 
1966), hostile especially to the mechanical application of apriorisms; G. Durand, Les 
structures anthropologiques de I'imaginaire (Paris 1973); Sabbatucci 1978, esp. the eponymous 
chap. 9; G. Ferraro, 77 linguaggio del mito. Valori simholici e realta sociak nelle mitologie primi- 
tive (Milan, 1979); M. Detienne, The Creation of Mythology (Chicago, 1986, original ed., 
1981); M. Godelier, La production des grands hommes (Paris 1982). I have learned nothing 
of value from the theories of Rene Girard, who, as is well known, understands myth 
as a system of persecutory representation, which functions to exorcise an aboriginal ill, 
comparable to the manner in which the scape-goat functions as a propitiatory victim: 
Qu'est-ce qu'un mythe? in: idem, Le Bouc emissaire (Paris 1983; English tr. as The Scape- 
goat [Baltimore 1986]) 37-67 (cf. the critique of his earlier theories by Gordon 1979b, 
279-310). Other points of view: H. Limet and J. Ries (eds.), Le mythe, son langage et son 
message. Actes du Collogue de Liege et Louvain-la-Neuve, 1981, Homo religiosus 9 (Louvain 
1983) is interesting as a methodological application; instructive too is the collection of 
articles in A. Dundes (ed.), Sacred Narrative. Readings in the Theory of Myth (Berkeley 1984); 
see too B. Lincoln, Myth, Cosmos and Society: Indo-European Themes of Creation and Destruction 
(Cambridge, Mass. and London 1986), 41ff, where he analyses the relations between 
myth and the social order. Of the Paris school with its Levi-Straussian and Dumezilian 
inflection, note J. -P. Vernant, Myth and Society in Ancient Greece (Hassocks 1980, original 
ed., Paris 1974); also the fine book by R.G.A. Buxton, Imaginary Greece: The Contexts of 
Mythology (Cambridge 1994). P. Leveque, Betes, dieux et hommes. L'imaginaire des premieres 
religions (Paris 1985), 117ff is unforgettable. Further: AAVV. Le mythe et le mythique. 



20 CHAPTER ONE 

In mediating complex symbolic structures, myth thus provides 
order, imposes meaning on reality; it offers the members of the group 
a simple means of self-identification, the collective acceptance of a 
system of beliefs. However, at the same time, the socio-cultural life of 
the community needs more or less explicit regulation. It is at this level 
that myth plays a part, since it tacitly mediates the ethical norms that 
condition and justify divine behaviour. Then again, membership of the 
cultural group in which such a system has developed demands a type 
of group behaviour capable of reproducing the mythical account in a 
sacralised, symbolic form, and in a regular manner, such that, by iter- 
ating the temporal rhythms and the order of things that the collective 
imagination has wrested from chaos, it can transmute them into the 
foundations of its own account of reality. 

Myth can of course be understood at several different levels. From 
my perspective, however, it is best understood as a more or less coher- 
ent explanatory account of 1) the heavenly bodies and the forces of 
nature, that is, the cosmic order; 4 2) the relation between human beings 
and their social environment, that is, the social and productive order; 
and lastly 3) a story about the after-life, whether in the underworld 
or in heaven — that is, the eschatological order. These three 'orders of 
things', cosmic, social and eschatological, make up the sub-system of 
beliefs which, together with that of values, i.e. ethical norms (whether 
codified in law or not), constitutes the religious system. 



Collogue de Cerisy-la-Salle (Paris 1987), esp. the outstanding contribution by J. Durand, 
reflecting on historical versus mythical change. For a survey of the various currents of 
thought in the field, beginning with the Presocratics, note J. de Vries, Forschungsgeschichte 
der Mythologie (Munich 1961); on classic twentieth-century theories: I. Strenski, Four 
Theories of Myth in the Twentieth Century: Cassirer, Eliade, Levi-Strauss and Malinowski (Iowa 
City 1987, 1989 2 ); J. Mohn, Mythostheorien (Munich 1998); and the fine survey by Csapo 
2005. I have also read with pleasure the critique by A. Testart, Des mythes aux croyances. 
Esquisse d'un theorie generate (Paris 1991); his 'Structure S' is refreshing. 

1 On which see e.g. R. Gothoni and J. Pentikainen (eds.), Mythology and Cosmic Order, 
Studia Fennica 32 (Helsinki 1987). I would also recommend the suggestive work of 
J.-R Cebe, Chaos et cosmos dans les civilisations traditionelles et antiques: le mythe 
et ses prolongements, Cahiers des Etudes Anciennes 28 (1993) 111-122, who offers a gen- 
eral account of the opposition between the two terms and their symbolic projection. 
This topic continues to be of concern to scholars interested in the universe viewed in 
religious terms; on the Renaissance Latin hymns of the Greek exile Michael Marullus 
(1453—1500), for example, see C. Harrauer, Kosmos und Mythos. Die Weltgotthymnen und die 
mythologischen Hymnen des Michael Marullus. Text, Ubersetzung und Kommentar, Wiener Studien, 
Beiheft 21 (Vienna 1994). On a different aspect, though one less directly relevant to 
my topic here, arising from current anthropological views, see W.C. Roof (ed.), World 
Order and Religion (Albany 1991). 



RELIGION, CULT AND MYSTERY 2 1 

This brief presentation of my pre-conceptions should enable the 
reader to understand my view of the controversial subject of the reli- 
gious character of the oriental cults. It was once wrongly believed that 
they were foreign 'national' religions, that is, state cults in their countries 
of origin. It is now a commonplace to claim that in fact all three of the 
cults I deal were profoundly transformed by contact with Graeco-Roman 
culture. 3 As cults integrated into the religious culture of the Latin West, 
they formed part of the ideological superstructure. That is, they are 
best seen not as alternative religious systems but as specialised formations 
within a larger cultural whole, where each individual could find his own 
place in accordance with his birth and disposition. 6 The context of that 
individual search however was always membership, whether voluntary 
or enforced, in the public or civic cult, a membership expressed both 
in personal piety and through participation in euergetistic public sacri- 
fice. 7 In such circumstances, it is reasonable to say that, at any rate by 
the second century AD, the adherents of these cults were in no sense 
trying to break free of their given world but invoking deities that were 
seen as powerful partners auxiliary to the system. 8 

That said, it seems to me that the novelty of the narrative-complexes 
that sustained these cults needs greater emphasis than is usually given 
them. That has been one of my main aims in writing this book. These 
complexes grounded a coherent set of beliefs, values and ritual practices 
that were very largely independent of the grand mythological structure 
that sustained Graeco-Roman civic cult in the wide sense (witness the 
central role of myth for the Second Sophistic). To that extent, I think, 
they can be seen as breaking the mould of ordinary religious practice 
in established groups, whether thiasoi or collegia. 9 On the basis of these 
myth-complexes, the oriental cults constructed their own distinctive 
religious cultures. In the case of the Mater Magna and Isis, they did 



5 A good general idea of the changes undergone by local cults, including those which 
did not succeed in spreading, is given byJ.Z. Smith, Native Cults in the Hellenistic 
Period, History of Religions 11 (1971) 236-47. 

6 Dowden 2000 is an interesting introduction to the way pagan religions worked. 

' See Gordon 1990b, 1990c; also the stimulating contribution by C. Ando, Exporting 
Roman Religion, in Rupke 2007d, 429—445. For an indispensable account of personal 
piety and its evolution in the Greek world to the Roman imperial period, see the rich 
study by Versnel 1981a. 

8 The problem is in fact a false one, as J.A. North, Continuity and Change in Roman 
Religion, PBSR 44 (1976) 1-12; Liebeschuetz 1979; 1992; and K. Hopkins, Death and 
Renewal (Cambridge, 1983) 82f, have clearly shown. See now Leppin 2007, 97-100. 

9 See recently the essays in Rupke 2007c. 



22 CHAPTER ONE 

so, to a significant degree, through the recognised means of festival and 
procession. But they also created their religious identity through appeal 
to the Greek category of mystery-cult. Indeed, so far as we know, the 
cult of Mithras existed only in this form. In taking over this mode of 
religious praxis, they extended the existing category of 'mystery' beyond 
its normative forms at Eleusis and Samothrace, the Bacchic mysteries 
and the Cabirion. For that reason, I see the truly characteristic feature 
of the oriental cults as residing in their provision of specific experi- 
ences of initiation, whose precise contents evidently in some cases, if 
we can trust Lucius' account in Apuleius, Met. 1 1 , were considered to 
be the private property of individual groups. 10 In the course of this 
book, therefore, although I usually refer to them as the oriental cults, I 
sometimes, where the argument requires special emphasis on initiation, 
its pre-suppositions and its consequences, use the term 'mysteries' or a 
periphrasis such as 'mystery component'. 11 

Hardly any one would go so far as to compare the oriental cults 
with sects in the properly sociological sense. 12 One can however in my 
view go too far in the opposite direction: Walter Burkert for example 
famously compares ancient initiation into the mysteries to the experi- 
ence of pilgrimage to Santiago di Compostela. 13 1 would go along with 
the traditional modern view, which sees mystery-cults as typologically 
different from public cult, and therefore in a certain tension with it, 
despite the fact that the celebrations at Eleusis and in many other 
places were an aspect of public or official religion. I should also men- 
tion that Giulia Sfameni Gasparro has gone so far as to suggest that 



10 The Isiac authorities at Rome did not recognise the initiation Lucius had undergone 
at Cenchreae and demanded that he undergo another, cf. Burkert 1987, 45. 

11 G. Sfameni Gasparro observes that in the first edition of the book I seemed 
systematically to confuse the term 'oriental cults' with 'mysteries' (Sfameni Gasparro 
2006). I evidently failed then to make my use of these terms clear, and am now aware 
that I often used 'mystery' too loosely. I have tried in this edition to make the matter 
clearer. 

12 Cf. e.g. R. Towler, Homo Religiosus: Sociological Problems in the Study of Religion (London 
1974) 110-15. But note e.g. Freyburger-Galland 1986. 

13 Burkert 1987, 10. No doubt he was thinking here of an experience of pilgrim- 
age analogous to that developed by V. and E. Turner, Image and Pilgrimage in Christian 
Culture. Anthropological Perspectives (New York 1978). (In that connection, note the inter- 
esting remarks of R. Gothoni, Pilgrimage = Transformational Journey, in Ahlback 
1993, 101—15. Understanding pilgrimage as a 'journey of transformation', Gothoni 
deliberately blurs the specificity of initiation, so that the journey (the preparation) 
converges with the rite of passage itself). Turcan 1992a seems to take a similar view 
to Burkert's, albeit less explicitly. Much the same is true of most essays in Freyburger- 
Galland 1986, despite the title. 



RELIGION, CULT AND MYSTERY 23 

only the cult of Mithras can be counted a true 'mystery religion'. 14 This 
seems to me, however, an unnecessary refinement, and I maintain the 
usual practice of taking all three as different instances, united by their 
ultimate origin outside the Graeco-Roman world, of the wider class 
of mystery-cults. 

I turn now to an analysis of the myths for their information about 
the different sub-systems I have mentioned. 



11 Sfameni Gasparro 1979c and 1985, xvi. She argues substantially the same view at 
greater length, but with different arguments, in her rather ambivalent piece: I misteri 
di Mithra: Religione o culto?, in Hinnells 1994, 93-102. 



CHAPTER TWO 
SYSTEMS OF BELIEF 



a7iep|ia kockiok;. . .ta [ivaxr\pia. 

Clement of Alexandria, Protrepticus 2.13.5 
Stahelin, p. 22.20f. Marcovich 

In his Exhortation to the Greeks, Clement of Alexandria claims that the 
myths of the deities worshipped in the mysteries are the seed of evil 
and corruption. The modern historian however sees them rather as 
synthesising a particular version of reality that formed the basis of 
the beliefs of the adherents of these religious systems. Thanks to their 
ability to integrate contradictions and cultural inconsistencies, as well 
as the processes of transformation each cult had undergone on its way 
from its point of origin to its entry into the ideological superstructure 
of Rome, these narratives formed more or less highly-developed organic 
wholes. 1 However the lacunate nature of the sources available to us 
means that our access to each complex is fragmentary and partial. Not 
merely have we lost the evidence relating to the intermediate stages 
of their evolution, we have in some cases, notably Mithras, lost even 
the mythical narrative itself. Fortunately myth was not the only means 
of instructing initiands, since the lived experience of the religious 
group provided many other means of transmitting the relevant reli- 
gious culture, to which we have some access. Fundamentally, however, 
the main information at our disposal for reconstructing the universe 
of beliefs is the mythical narrative, where it survives, and, to a lesser 
extent, the iconography together with references in literature, and of 
course epigraphy. 

The reconstruction we might call 'canonical', with an account of the 
content of each cult, is of course that of Franz Cumont, dating from 
1 905-6. 2 As I have pointed out already in the Introduction, subsequent 



1 I naturally accept the point made by Bourdieu 2001, 338: "II ne faut pas, on le 
voit, demander au mythe, meme 'rationalise', plus de loqique qu'il n'en peut offrir". 

2 The first edition of Cumont 1929 appeared in 1906; second revised ed., 1909. The 
book was very soon translated into English by G. Showerman (191 1), himself the author 
of a book on Cybele: The Great Mother oj the Gods (Wisconsin 1901); this translation, of 
the second French edition, was re-issued in 1956 by Dover, and reprinted frequently 



26 CHAPTER TWO 

research has revised or questioned almost all of his assumptions, which 
were themselves largely taken over from his immediate predecessors 
writing in French, especially Ernest Renan and Victor Duruy. Since I 
do not intend here to provide a proper survey of the literature since 
Cumont, I will restrict myself to providing a summary account of it. 3 

I have already made clear that in my view the most characteristic 
typological feature of the oriental cults was the promise of salvation, or 
rather, more precisely, their deities' power to overcome Fate. I need to 
express myself carefully here, since over the years many scholars have 
questioned this crucial element of the old grand narrative. 4 The objec- 
tion is indeed partly justified, since I freely admit that mistakes were 
made in interpreting the characteristics of these gods by assigning to 
them features documented in Christianity but not in the oriental cults. 
In fact, quite apart from the borrowings and transfers between pagan- 
ism and Christianity in antiquity, modern scholarship, in making use of 
information outside the strict limits of its applicability, has encouraged 
the development of what we might call an 'interpretative osmosis'. 

We can thus no longer confidently claim that the mystery-component 
of the oriental cults offered their adherents the hope of resurrection, or 
the firm promise of eternal life in the company of the god. s In older 



thereafter, most recently in 2003. (The third edition of 1928 was simply a reprint of 
the second: Bongard-Levine et al. 2007, 87 n. 147. The much superior 4th ed. has 
never been translated into English; a 5th French edition appeared in 2006 as part of the 
systematic republication of Cumont's works; see Bonnet and van Haeperen 2006.) 

3 Several articles in ANRW provide bibliographic surveys of this process of revision 
and re-interpretation, e.g. B.M. Metzger, A Classified Bibliography of the Graeco- 
Roman Mystery Religions, 1924-1977, ANRW 11,17.3 (1984) 1259-1423; Beck 1984 
and 2004b. It is still worth consulting the older surveys by Pettazzoni 1924/1997 and 
Metzger 1955. 

1 Doubtless following the scepticism of Wilamowitz 1931—2 regarding the soteriologi- 
cal promises of these cults, Prumm 1943/1954, 313ff and then R. Bultmann, Primitive 
Christianity in its Contemporary Setting. Transl. R. Fuller (London and New York 1956) [orig. 
1949], 173 rejected the idea that Mithras was a saviour-god in the Cumontian sense. A 
generation later, both Brelich 1965 and Frankfort 1958 were likewise sceptical of the 
entire soteriological scenario; the sharpest recent critics have been MacMullen 1981 
and Burkert 1987: 66—88; cf. the doubts of Sfameni Gasparro 1985 regarding Cybele 
and Attis. The power of the mystery gods to overcome Fate — admittedly the case of 
Mithras is an at least partial exception — is however not questioned by the author of 
the most exhaustive study of the topic, Magris 1985, 2: 505ff We shall have occasion 
to return to this difficult issue later. 

: ' It is precisely on these general questions that subsequent scholarship has distanced 
itself most sharply from Cumont's assumptions. Nevertheless I think that this has 
involved the rather questionable triumph of minimalist positions, hypercritical in rela- 



SYSTEMS OF BELIEF 27 

scholarship, the redeeming, healing and salvific character of these 
cults was nudged in the direction of what is understood as 'salvation' 
from a Christian perspective, namely evading the malignity of Fate. 
But now that such a marked dualism in relation to post-mortem fate 
has been rejected, it is far more difficult to decide quite what the offer 
of salvation referred to. 6 It would suffice if it were simply a matter of 
depriving Fate of its inexorability, both in the real world and in some 
imaginary universe, that is, in both the 'locative' and the 'utopian' 
dimension,' in such a way that the threat posed to the Graeco-Roman 
cultural system by the existential unpredictability of daily life would 
be fundamentally allayed. Even so, however, it is difficult precisely to 
define the offer and its effects. 

The words Cicero puts into his own mouth in his dialogue De legibus, 
which cannot possibly be suspected of Christian contamination, pro- 
vide us with one view of the claims and significance of the mysteries. 
Although he refers specifically to Eleusis, the point made holds good 
for the other mysteries that took Eleusis as their model: 

Marcus: Nam mihi cum multa eximia divinaque videntur Athenae tuae [a reference 
to his interlocutor, T. Pomponius Atticus] peperisse atque in vitam hominum 
attulisse, turn nihil melius Mis mysteriis, quibus ex agresti immanique vita exculti 
ad humanitatem et mitigati sumus, initiaque appellantur ita re vera principia vitae 
cognovimus, neque solum cum laetitia vivendi rationem accepimus, sed etiam cum spe 
meliore moriendi. s 

Cicero, De leg. 2.36 



tion to the evidence and little disposed towards moderate solutions. I do not subscribe 
to the maxim in medio Veritas, since the truth may lie far from the mid-point between 
two scholars; there may always be a third, a fourth, an nth possibility ... I therefore do 
not claim personally to prefer easy eclectic positions. 

6 We have moved from a historiographical situation in which salvation was conceived 
solely in eschatological terms to one in which the sole admissible content is intramun- 
dane. I have no doubt that the reality was much more ambiguous, rich and subtle, so that 
we need to think in terms of a flexible model. I prefer to leave the question undecided 
here and come back to it in section 3 of this chapter, on the world beyond. 

7 I adopt the terms introduced by J.Z. Smith 1990, 121-42. 

8 "It seems to me that, among the many admirable and divine things you Athe- 
nians have established to the advantage of human society, there is nothing better than 
the mysteries, by means of which we have been polished and softened into civilised 
behaviour out of the austerities of barbarism. They are justly called initiations, for it 
is mainly through them that we learn the basic principles of living, not only the art 
of taking pleasure in life but also of dying with greater hope." 



28 CHAPTER TWO 

In our case, victory over Fate is confirmed by the texts we possess 
concerning the oriental cults. The clearest formulation occurs in the 
'Self-predication of Isis' texts, where the goddess Isis claims to be 
victorious over Fate: 'Eyro to eirjapiaevov vikco. 'E|a,(xi> to ei|acxp|jivov 
aKovei, 'I overcome Fate; Fate hearkens to me'. 9 Nor did Serapis 
lag behind her, for in the account of one of his miracles we find the 
expression: [xacA Moipaq yap eyro n£Tarj(pid^ro, 'I change the garb of 
[the] Moirai', 10 and an oracle of Apollo, presumably from Miletus or 
Didyma, says of him: avxbq yap iiouvoi; Kai Tat; Moipa; neOoSevei, 
'He alone even gets round the Moirai'. 11 Dozens of people could attest 
to his individual saving acts: |ieoTai Se ayopai . . . Kai A,i|aive<; Kai toc 
evptixropa tcov 7i6A,erov tcov Ka0' EKaaTa e^riyovnevrov. 12 However the 
most significant text relating to this issue is a passage from Apuleius' 
Metamorphoses: 

Vives autem beatus, vives in mea tutela gloriosus, et cum spatium saeculi tui permen- 
sus ad inferos demearis, ibi quoque in ipso subterraneo semirotundo me, quam vides, 
Acherontis tenebris interlucentem Stygiisque penetralibus regnantem, campos Eysios 
incolens ipse, tibi propitiam frequens adorabis. Quodsi sedulis obsequiis et religiosis 



9 Totti no. 1 §55f., cf. 46f. (tr. R.M. Grant). Totti also conveniently provides the 
references to the four known versions, most importantly that from Kyme = I.Kyme 
no. 41 (where Engelmann sets out the stemma, p. 101) = R1CIS 302/0204 = Sanzi 
2003, 213—5 no. 46. See further nn. 289 and 290 below; W. Peek, Der hishymnus von 
Andros (Berlin 1930) 120ff; D. Miiller, Agypten und die griechischen Isis-Aretalogien (Berlin 
1961) 74—85; Griffiths 1975, 243; also J. Bergman, 'I overcome Fate, Fate hearkens to 
me', in H. Ringgren (ed.), Fatalistic Beliefs in Religion, Folklore and Literature: Papers read at 
the Symposium on Fatalistic Beliefs held at Abo on 7th-8th September 1964 (Stockholm 1967) 
35—51, where he shows that Isis' claim to be Mistress of Fate was derived from Amon, 
the Egyptian creator-god, and that it was mainly understood to mean overcoming 
misfortune. Griffiths 1991, 323 bluntly asserts that the term salus, salvation, in the cult 
of Isis connotes a secure and protected life in this world and the next. 

10 PBerol. 10525 = D.L. Page, Select Papyri 3 (Loeb ed.) 424-29, no. 96 = Totti no. 
12, 1. 13 (IIP). The supplement is by L. Fahz. 

11 The oracle is preserved only in Cod. Vindob. 130 and Cod. Laur. 37; the addressee 
is Timaenetus (probably II P ). The text is most conveniently to be found as Totti no. 60. 
On the later oracles of Apollo, see now A. Busine, Paroles d'Apollon: pratiques et traditions 
oraculaires dans lAntiquite tardive. RGRW 156 (Leyden 2005), though, as far as I can see, 
she seems to have missed this one. However she does list two other oracles from Apollo 
at Didyma that tell the enquirer that he may pray to Serapis: I.Milet. 1.7, 205b — her 
cat. no. 43 (from the Serapeum in Miletus), pp. 59, 97. 

12 Aelius Aristides, Or. 45.30; cf. also the claim that Serapis has raised individuals 
from the dead (ot>TO<; Keiitevotx; avecrcrio'ev), ibid. 29. 



SYSTEMS OF BELIEF 29 

ministeriis et tenacibus castimoniis numen nostrum promeruis, scies ultra statutafato 
tuo spatia vitam quoque tibi prorogare mihi tantum licere. u 

Apuleius, Met. 11.6 

I believe this to be quite characteristic. 14 The goddess promises a blessed 
existence in the world beyond, after a life-span prolonged beyond the 
limits set by Fate, thus combining the two forms of salvation, physi- 
cal and spiritual. The text is thus a marvellous precis of what we are 
entitled to see as a sort of Isiac creed, or at any rate a sketch of one. 
In section 2.3 of this Chapter, on the World Beyond, I discuss how 
far the other cults aside from that of Isis shared in the soteriological 
promise, however we are to understand it, that in my view is central 
to these mystery-cults. I accept that they differed in the degree of 
emphasis they placed on life in the other world; but would claim that 
on the whole they laid more stress on it than traditional paganism, and 
in that respect were somewhat closer to Christianity 10 

Before concluding this introductory section it is worth thinking about 
how the mystery-component of the oriental cults has been understood 
as a religious phenomenon. What we might call Christian 'contamina- 
tion' is by no means solely a feature of the evidence from Antiquity. It 
re-appears in modern interpretation. An obvious example is this passage 
from Mircea Eliade, acceptance of which would require considerable 
prior contextualisation: 

For the history of religion, the particular importance of the Greco-ori- 
ental mysteries lies in the fact that they illustrate the need for a personal 



13 "Moreover you will live in happiness, you will live in glory, under my guardianship. 
And when you have completed your life's span and travel down to the dead, there too, 
even in the hemisphere under the earth, you will find me, whom you see now, shining 
among the shades of Acheron, and holding court in the deep recesses of the Styx, and 
while you dwell in the Elysian Fields I will favour you and you will constantly worship 
me. But if by assiduous obedience, worshipful service, and determined celibacy you 
win the favour of my godhead, you will know that I — and I alone — can even prolong 
your life beyond the limits determined by your fate" (tr. Hanson). 

14 For my views on how we are to read this novel, see n. 278 below, and Chap. 4 
n. 527. On the roles of fate and providence in the novel, see N. Fick-Nicole, De fati 
et providentiae et fortunae ratione, quae inter Apulei Metamorphoseon libros inter- 
cedit, in J. Blansdorf (ed.), Loquela vivida: donum nataliciumN. Sallmann oblatum (Wiirzburg 
1999) 59-73. 

15 Liebeschuetz 1992, 251. Liebeschuetz' is a considered view, one that seeks to 
discount positions which, consciously or unconsciously, take Christianity as their point 
of reference. 



30 CHAPTER TWO 

religious experience engaging man's entire existence, that is, to use Chris- 
tian terminology, as including his 'salvation' in eternity. 16 

It is unclear whether Eliacle understands the need for a personal reli- 
gious experience here as something inherent in all human beings, or 
just those of antiquity; 17 and the matter is further confused by the use 
of the Christian term 'salvation'. Unless such a claim refers to a specific 
historical period at which such a personal need became manifest, it is 
difficult to accept, and I remain sceptical despite the fact that Eliade's 
position would be convenient for my argument. Enough of methodol- 
ogy, however: I must resume the thread of the argument. 

The relationship between Christianity and the oriental cults has 
provoked much debate. I deal with some specific issues later on, and 
conclude the book with a chapter devoted to the historical — and histo- 
riographical — relation between the two. This is however the appropriate 
moment to raise a basic conceptual problem. Christian monotheism 
is hostile to the position of restrained tolerance that characterises the 
contemporary academy. In order to understand this intolerance, the 
academy investigates the origins of Christian monotheism, but this very 
research, paradoxically enough, simply adds fuel to the flames. 

Whereas monotheism is for some the ground of the authentic reli- 
gious feeling that sets Christianity at the top of the tree of religious 
evolution, which is tantamount to claiming its intrinsic superiority, oth- 
ers, emphasising the deep-rooted monotheistic tendencies in the Empire, 
reject the claim that monotheism is original to Christianity. In support 
of their thesis, they sometimes refer to the adherents of the mystery 
religions. In my view, there are two relevant considerations here. One 
concerns the very process of constructing Christian exclusivism, which 
I will not go into here, but which we need to bear in mind if we are 
to maintain a distance between religious discourse and its practice, as 
well as the attitudes of the congregations and their leaders towards 



"' M. Eliade, Rites and Symbols of Initiation: the Mysteries of Birth and Rebirth (New York 
1958) 113. 

17 Cf. Burkert 1996, a programmatic book in which he claims "the existence of 
biological patterns of actions, reactions, and feelings activated and re-elaborated 
through ritual practice and verbalised teachings, with anxiety playing a foremost 
role ... Religion follows in the tracks of biology". Another form of the naturalising 
argument is proposed by P. Boyer, The Naturalness of Religious Ideas: A Cognitive Theory of 
Religion (Berkeley and Los Angeles 1994). 



SYSTEMS OF BELIEF 3 1 

aberrations. 18 The second, which is more directly relevant here, concerns 
the issue of exclusivity, and the syncretistic tendencies that have been 
descried during the Imperial period. 19 

Despite its frequent repetition, I have little faith in the idea that there 
was a process of monotheistic syncretism that culminated in Christi- 
anity 20 In fact, Christianity found itself obliged to accommodate the 
polytheism that prevailed among the majority of the population by 
sacralising a variety of inferior beings and so created, under the aggres- 
sive banner of monotheism, an authentic, hierarchically-structured, 
Christian pantheon. Viewed in that light, the tendency to syncretise is 
not an historical reality at all, or at least not one that necessarily moves 
in a 'monotheistic' direction. A quite separate issue is whether there was 
a political will within paganism to integrate as many cults as possible 
into an hierarchical system intended to enhance the credibility and the 
interests of the imperial centre. At any rate, it is quite impossible to 
claim, on the basis of certain formulae of adoration characteristic of 
the oriental cults, that they too contributed to a some sort of mono- 
theistic imperative. These expressions in fact arise from a philosophical 
stance and had little impact on religious practice, however much the 
triumph of Christianity has made people think that the onward march 
of monotheism was irresistible. 

As the lucid discussion by Henk Versnel has shown, the best image 
for this phenomenon is the concept of 'henotheism' (1990, 35-38 et 
passim). 21 The term, which is modern, is calqued upon the formula 



18 The point concerns the process of giving formal expression to Christian exclusiv- 
ism, which of course only became sharp-edged once it became meshed with political 
power. In any case, anyone interested in how Late -Antique Christian exclusivism 
expressed itself, and the sort of things that were done in its name, should consult Trom- 
bley 1993-4; E. Sauer, The Archaeology of Religious Hatred in the Roman and Early Medieval 
World (Stroud and Charleston SC 2003); or Hahn 2006. Some pertinent reflections 
on pagan reactions in a lecture at the American Academy, Rome on 3 May 2007 by 
S. Anghel, Burying the Gods: Hiding Statues from Christians in Late Antiquity. 

19 The concept of syncretism is a sort of joker, used to excess by historians of 
ancient religion. Its ambiguity renders it extremely useful to those unwilling to define 
their terms very precisely. I prefer to use it cautiously; see the remarks by C. Stewart 
and R. Shaw (eds.), Syncretism /Anti-syncretism: The Politics of Religious Synthesis (London 
and New York 1994) 1-26. On Cumont's loose, quite untheorised, use of the term, 
note Motte 1999. 

211 Cf. the welcome criticism of the idea that the epithet v\\iiaxoq implies a pagan 
tendency towards monotheism by N. Belayche, Hypsistos. Une voie de l'exaltation des 
dieux dans le polytheisme greco-romain, ARG 7 (2005) 34—55. 

21 The concept has unfortunately not become widely familiar (there is no lemma 
either in the Oxford Classical Dictionary 3 or Der Meue Pauly). T. Kotula, Les Apologistes 



32 CHAPTER TWO 

eiq 6 9eoq (ordinarily translated 'one is the god' but in fact meaning: 
'this/our god is no. 1'), which implies not exclusivism but a more or 
less marked preference on the dedicator's part. It bears repeating that 
henotheism, despite its relative frequency in our sources, was not a 
widespread form of belief in the Empire. It is also worth pointing out 
that the adherents of the oriental cults similarly formed an insignifi- 
cant minority within the overall population, their importance inversely 



africains du IIP siecle face aux tendances monotheistes pa'iennes, Histoire et archeologie 
de I'Afrique du Mord. Spectacles, vie portuaire, religion. Actes du V Colloque international (Paris 
1992) 153-58 uses it appropriately, as does Turcan 2000, 145-52 ('L'henotheisme 
mithriaque'). Recent publications display a surprising ignorance of Versnel's work and 
continue uncritically to insist on 'monotheistic tendencies'. I think the reason must be 
that the term 'monotheism' is believed to have more attraction than 'henotheism'. If 
one looks at the most recent contribution to the topic, Athanassiadi and Frede 1999, 
one is astonished at the confusions that the contributors land themselves in by talking 
constantly about 'monotheism' in a polytheistic context, especially since they man- 
age to believe that monotheism was an inescapable tendency. The term 'henotheism' 
only occurs once (M.L. West, Towards Monotheism, in Athanassiadi and Frede 1999, 
21-40), yet both he and his co-contributors would have found it much more useful 
than the fancy label given the seminar which in turn gave its name to the volume. 
And given that West describes the concept perfectly ("Where we see a god emerging 
as plenipotentiary, the existence of other gods is not denied, but they are reduced in 
importance or status, and he is praised as the greatest among them", p. 24), it is even 
more extraordinary that they continue to talk about 'monotheistic tendencies'; and 
that even Wolf Liebeschuetz squeezes his foot into the same shoe, when his conclusion 
is rather that we have to do with a sort of personal syncretism, especially among the 
Neo-Platonists, though in practice they were all polytheists (Liebeschuetz, 1999). We 
unfortunately encounter the same terminological error in Digeser 2000, aggravated 
by the fact that, dealing only with a literary text, she universalises the movement 
in favour of monotheism. The truth is, I think, quite different: it was not paganism 
that became monotheist, but Christianity that became polytheist. It was thus that 
the Church in the end required the faithful to venerate the Virgin, the angels, and 
the saints as minor divinities, as substitutes for their now abandoned manifold divine 
beings. Christian monotheism is actually nothing but a dogmatic claim, formally 
accepted by Christians — despite the utter illogicality of the idea of the Trinity — who, 
however, continued to behave in practice exactly like the adherents of pagan polythe- 
ism [a remark that scarcely applies to Protestants, however!]. Moreover, the project of 
discovering monotheism in the oriental cults, suggested by the editors as the subject 
of another seminar ("So crucial an area of pagan monotheism as the theology of the 
mystery cults has not been touched on in this volume; it is our intention to examine 
this important theme in a future seminar", ibid. p. 20 n. 49) is perfectly de trop. If they 
had bothered to open G. Fowden, Empire to Commonwealth. Consequences of Monotheism in 
Late Antiquity (Princeton 1993), or Bradley 1998, their planning might have been more 
to the point. Fowden, basically following Versnel, summarises the differences between 
the concepts polytheism, henotheism and monotheism (p. 5), and rejects the tendency 
to extend monotheism into the semantic range of henotheism (e.g. p. 40f); cf. too the 
discussion of Julian's Mithraism (pp. 52-56). As for Bradley, he understands Lucius' 
transformation not as a conversion but as an example of henotheism (1998, 331). 



SYSTEMS OF BELIEF 33 

proportional to their number. 22 No doubt economic factors, as well 
as the desire to assert their beliefs, should be invoked to explain their 
prominence in the epigraphic and monumental evidence. 

In the following section, I first set out some of the basic premises that 
define the relations between the oriental deities and their adherents; 
the remainder examines each of the myths successively in some detail, 
teasing out the major themes. 

1. Cosmic Order and the Nature of the Divine 

omnis enim per se divum natura necessest, 
immortali aevo summa cum pace fruatur 
semota ab nostris rebus seiunctaque longe. 
Nam privata dolore omni, privata periclis, 
ipsa suis pollens opibus, nil indiga nostri, 
nee bene promeritis capitur neque tangitur iraP 

Lucretius, de rerum natura 2, 646-52 

It is typical of the gods of the oriental cults that they have some expe- 
rience of human existence characterised by direct contact with death. 
Some indeed suffered it themselves, which would be unthinkable for 
the Olympian gods, whose manifold experiences do not include their 
own deaths. Only a rhetorician such as Lactantius in the early fourth 
century AD could lump all these deities together: 

Let us please consider the anguish of the gods who were unlucky. Isis lost 
her son, and Ceres her daughter; Latona was driven out and harried all 
over the world, only with difficulty finding a little island to give birth on. 
(7) The mother of the gods fell in love with a pretty youth, and when she 
caught him with a paramour, she castrated him and made him a eunuch, 
and that is why his ritual is celebrated even now by the priests called Galli. 
Juno persecuted her brother's paramours so fiercely because she could 
not get pregnant herself ... (9) The indecency of Venus is beyond words, 



22 Since it is not my intention to survey either the expansion or the sociology of the 
oriental cults (this will be one of the topics of a projected book on the mysteries in 
Hispania), I may here cite three such studies relating to our deities: Schillinger 1979; 
Mora 1990; and Clauss 1992. For IOM Dolichenus, see E. Sanzi, Dimension sociale et 
organisation du culte dolichenien, in G.M. Bellelli and U. Bianchi (eds.), Orientalia Sacra 
Urbis Romae: Dolichena et Heliopolitana. Studia Archaeologica 84 (Rome 1996) 477-513. 

2i "For it is essential to the very nature of deity that it should enjoy immortal exis- 
tence in utter tranquillity, aloof and detached from our affairs. It is free from all pain 
and peril, strong in its own resources, exempt from any need of us, indifferent to our 
merits and immune from anger" (tr. R.E. Latham). 



34 CHAPTER TWO 

prostituting herself to the lust of one and all, not only gods but also men. 
She it was who bore Harmonia as a result of Mars' famous rape of her, 
and by Mercury she had Hermaphroditus, who was born bi-sexual. By 
Jupiter she had Cupid, by Anchises Aeneas, by Butes Eryx; only with 
Adonis did she fail, because he was gored to death by a boar when still 
a boy. (10) As it says in the Sacred History, she started prostitution, and 
promoted it on Cyprus as a way the women could make money from 
public hire of their bodies: she required it of them to avoid herself being 
seen as the only wicked woman, with a gross appetite for men. 

Lactantius, Inst. div. 1.17.6—10 (tr. A. Bowen) 

Despite this rhetorical invective, which carries on to take in other dei- 
ties too, it is quite clear that death is almost completely foreign to the 
gods of the traditional pantheon, and that Lactantius' bundling them 
all together is a travesty. Moreover, the mystery gods' direct experi- 
ence of death is fundamental to what they were subsequently able to 
achieve: life can triumph only because they have gained immortality. 
Death brings them close to human beings, while the rebirth they offer 
has a grandeur about it unattainable by the traditional gods of the 
Graeco-Roman pantheon. 

In the later nineteenth and early twentieth centuries, however, the 
History of Religions school and the Comparativists fostered a caste of 
mind that favoured collocations nearly as arbitrary as those of Lactan- 
tius. They asserted that the experience of death in the oriental cults 
was the same in each case: the gods suffered it physically but succeeded 
in overcoming it by virtue of their vital force, their power as creators, 
demiurges, and saviours. Osiris died by being cut up into pieces, but Isis 
managed to revive him. Cybele caused the death of Attis, but allowed 
him to carry on living. Though he did not suffer death himself, Mithras 
slaughtered the bull from which life arose anew. Death and immortality 
were interpreted as symbols of the annual regeneration of nature, which 
was understood to be the central focus of these religions. A century 
of debate has served to show that the reality was far more nuanced 
than it once seemed. 24 Osiris was resurrected but his resurrection was 
restricted to the Underworld; the very existence of Attis' resurrection 



21 Cf. "Mais ce qui demeure essentiel, au niveau des pratiques cultuelles (tout a fait 
differentes de ce que peut produire la speculation philosophique), c'est la diversite, 
la multiplicite, la distinction": P. Borgeaud, L'Histoire (comparee) des Religions: une 
discipline au futur, in G. Sfameni Gasparro (ed.), Themes and problems in the History of 
Religions in Contemporary Europe: Proceedings of the International Seminar, Messina March 30-31, 
2001 (Cosenza 2002) 67-77 at 76, on Mother or 'Great' goddesses. 



SYSTEMS OF BELIEF 35 

in any form has been seriously doubted; Mithras himself could never 
have had anything to do with resurrection at all. So much can be 
freely admitted. In my view, however, the triumph over Fate remains 
a constant; and, from a certain point in the High Empire, salvation in 
the other world came to be a deep conviction shared among many of 
the adherents of the oriental cults. The main reason for thinking this 
is that their central rituals, to which I shall later devote more particular 
attention, are in fact initiatory, and were replete with the symbolism 
of death and resurrection. 25 Since this symbolism is so transparent, it 
seems perverse to deny the centrality of the belief in these cults. 

Of course, it is not merely the fact that they have lived that defines 
these gods as mystery-divinities. There can be nothing more anthropo- 
morphic than the Homeric gods, with their enviable vices and virtues. 
However the most striking peculiarity of those traditional deities was 
that they had no share in one of the most private of human experi- 
ences, death. They were immortal. By contrast, the gods of the oriental 
cults shared with their adherents in one way or another the ultimate 
rite of passage, the transition from being to not-being 26 Thanks to this 
experience they acquired a special claim to be able to attend to the 
problems, anxieties and needs of human beings, so much so that these 
concerns are to all appearances the main preoccupations of the divine 
world. This was certainly the case in the first three centuries AD. 

The myths however are more overtly concerned with the establish- 
ment, or the re-establishment, of the order of the cosmos. In the case of 
the Egyptian cults, this order was established by the combat in heaven 
between Horus and Seth; Isis is consequently celebrated as mistress of 
the universe, exercising power over gods, humans and nature alike. 2 ' 
The same emphasis is also evident in Mithraic iconography, where the 
cult-relief offers a symbolic representation of this assertion of order. 28 



23 The myth recounted the experiences and suffering of these gods and was repro- 
duced in the ritual: Isis, for example, did not allow her sufferings to be engulfed in silence 
and oblivion, but "infused images, suggestions and representations of her experiences 
at that time" into the sacred rites: Plutarch, De hide 27, 361d-e. 

26 See the rather uneven Louvain collection (Ries 1986b), which however contains 
some interesting contributions on transition-rituals in the mysteries; cf. Bianchi 1986b. 
More discussion of this issue in Chap. 4. 2. a. 

27 See the inventory drawn up by Malaise 1986a, 27ff; cf. Mora, 1990, 2: 60f. 

28 The bibliography here is very large. For the moment I just need to cite Beck 
2004a and 2006. The theme continues to be of central interest in Mithraic studies, 
as is plain from the number of 'decodings' of the tauroctony-scene (on which see 
Beck 2004c). One example would be Ulansey 1989, who, whatever the merits of his 



36 CHAPTER TWO 

Although the case with the Phrygian cults is not quite so clear, there 
too the connection with the cosmos is undubitable. Both literary texts 
and iconography attest to the cosmic roles played by both Cybele 
and Attis. Thus Lucretius alludes to a cosmological interpretation of 
Cybele's lion-drawn chariot: aeris in spatio magnam pendere docentes tellurem 
neque posse in terra sistere terram, 29 that seems to be picked up by a coin 
of M. Plaetorius Cestianus as curule aedile for 68 or 67 BC, showing 
Cybele's head with a globe in front of it. 30 The evidence for Attis is 
more extensive. 31 Indeed, the process of turning him into a 'divine 



particular thesis, at any rate accepts the Roman cult as the product of an intellectual 
exercise by a small group of individuals at a specific point in time; see also Beck 1998a 
(= 2004a, 31-44). That may be so, as we shall have occasion to see; but we should 
not forget the point made by E. Will, ap. J. Duchesne-Guillemin, Sur l'origine des 
mysteres de Mithra, CRAI 1990, 281—4 at 285f., who observes that it is not theologians 
who create religions, but religions that make theologians. I incline to agree with him, 
notwithstanding the case of Serapis, which might be considered a counter-example. At 
the same time, one should note Beck's analysis of the relations between the initiatory 
grades and the planets as reproduced in the structure of the cult and the rituals (Beck 
1988, 7—11). Ulansey modifies Beck's scheme by identifying Mithras with the constel- 
lation Perseus, which makes no sense from the point of view of the religious content. 
Ulansey's 'extravagance' lies much more in his scarcely-qualified conclusion. 

29 De rerum nat. 2. 602f.: "Thus teaching that the great world is poised in the spacious 
air and that earth cannot rest on earth". The lions were later interpreted as pointing to 
an association between Cybele, now understood as universal Providence, and Helios: 
Julian, Or. 5. 167b; cf. Sfameni Gasparro 1981, 40 If 

30 Cf. Summers 1996, 341. For Cestianus see F. Munzer, s.v. Plaetorius no. 16, RE 
20 (1950) 1950-52. I am not convinced however that the glass globe (if that is what 
it is) in the 'Cybele' (or Venus) fresco in the taberna of the house of M. Vecilius 
Verecundus at Pompeii (Reg. IX, ins. 7.1) represents the world ruled by Cybele and 
Attis (CCCA 4, 17f. no. 42). 

31 Frazer 1914, 1: 281ff, wrongly imagined that the key lay in the connection 
between Attis and Zeus, supposedly suggested by the epithet Papas (which obviously 
means 'Father') in the first of the so-called Hymns to Attis collected by the Christian 
polemicist pseudo-Hippolytus, in the context of his attack on the Naassenes (a gnostic 
group which had developed its own discursive account of the myth of Attis, cf. Borge- 
aud 2004, 102—7). Nevertheless the passage is important in relation to Attis' cosmic 
dimension. It reads: Kai oi f&pvyei; aXkoxt uev Ila7tav, noxk 8' <at>> veicuv rj 9eov 
r] xov aicapTtov f) ainoXov r\ xXoepov azaxov d|xr|0evTa, f\ <t>6v noXuKapnov exiK-cev 
a(j,vy8aA,oi;, dvepa copucrdv, And the Phrygians [name you] sometimes Papas, some- 
times Corpse, or God, or Fruitless, or Goat-herd, or "Already-harvested unripe Ear of 
grain", or "the fruitful One to whom the almond-tree gave birth", Male flute -player' 
(Refutatio 5.9.8 = Sanzi 2003, 287: Cybele no. 41). For, earlier in this section, [Hip- 
polytus] interprets the epiclesis Aipolos' (Goat-herd) as the equivalent of d<e>ut6Xo<;, 
'eternal sphere': 'he constantly moves (6 del rcoXSv) and revolves and carries round the 
entire cosmos by his circular motion' (5.8.34, omitted by Sanzi). See the analyses of 
M. Marcovich, The Naassene Psalm in [Hippolytus] , Haer. 5.10.2, in B. Layton (ed.), 
The Rediscovery of Gnosticism, 2 (Leyden 1981) 770-78 = idem, Studies in Graeco-Roman 
Religion and Gnosticism (Leyden 1988) 80—89; Sfameni Gasparro 1981, 405—08; Turcan 



SYSTEMS OF BELIEF 3 7 

being with cosmic and pantocratic prerogatives' is nowadays clearly 
recognised (Lancellotti 2002, 135). There are numerous references to 
it in the fourth century and later, such as the following passages from 
Sallustius, the friend of the emperor Julian, like him a proponent of 
the allegorical interpretation of myth and the mystical interpretation 
of cult-practice: 

The Mother of the gods is a life-giving goddess (^cooy6vo<;), and therefore 
she is called mother, while Attis is creator (8r]|n,o , upy6<;) of things that come 
into being and perish, and therefore he is said to have been found by the 
river Gallos: for Gallos suggests the Galaxios Kyklos (Milky Way). 

Sallustius, Llepi tcov Gecov kou ton koctiiod §4, 
p. 8 11. 3-7 Nock (tr. Nock) 32 
And again: 

The Mother loves Attis and gives him heavenly powers (signified by the 
cap [7uA,o<;]). 

ibid., p. 8 lines 8-10; cf. Julian, Mother 165b; 170d-171a 33 

Another explicit passage, Macrobius, Sat. 1.21.9, seems at first sight to 
be still later, but points us in fact to one of the main sources of this solar 
syncretism, Porphyry's llepi ayaXuttxcov, written in the third quarter 
of the third century AD. Macrobius writes: Solem vero sub nomine Attinis 
ornant fistula et virga. M The earlier date is confirmed by a passage from 
Arnobius, Adv. gent. 5.42.5, whose source must also lie at least as far 
back: Attidem cum nominamus, solem. . . significamus et dicimusP Porphyry is 
indeed one of the nodal points in the development of the syncretistic 
integration of quite diverse divinities into association with the Sun. 



1996b: 393-95; Lancellotti 2000: 272-84; eadem 2002, 115-18. H. Graillot claimed 
that Attis might also have functioned as a 'Father Heaven' over against Cybele's 'Mother 
Earth' (1912, 15), though the suggestion has never found favour. In my view, however, 
it could help to understand the significance of the myth in a wider sense, once the 
symbolisms had been merged. 

32 = Sanzi 2003, 295: Cybele no. 44; cf. Julian, Or 5. 165b-c; 166a-b. 

53 On the correspondences between Sallustius and Julian here, cf. G. Rochefort, 
Le riepl 0ecbv kcu kog|J,oi) de Saloustios et l'influence de l'empereur Julien, REG 69 
(1956) 50-72 at 61. 

34 — Sanzi 2003, 307: Cybele no. 53. "[The Phrygians] decorate the Sun, under the 
name of Attis, with a pan-pipe and a. pedum (shepherd's crook)." 

35 "When we name Attis... we mean and speak of the sun." On Arnobius' aims 
here, cf. Mora 1994, 190f Note also in the same sense Firmicus Maternus, De errore 
8.1-3 (probably also from Porphyry) and the late texts cited by Lancellotti 2002, 135. 
For her part, the Mater Magna becomes the mother and consort of Zeus-Jupiter, 
mistress of all life, the cause of all coming into being (Julian, Or. 5. 166a); cf. Lancel- 
lotti 2002, 128. 



38 CHAPTER TWO 

If Attis was drawn into this process of neo-Platonist syncretism, 
however, it was because he already had solar, and more generally cos- 
mic, associations: these more or less novel formulations contain ideas 
that originated much earlier than Porphyry, in a cosmology current 
within the cult of the Mater Magna. For the solar association of Attis 
appears in the visual evidence already in the second century AD, most 
strikingly in the well-known image of him dedicated by C. Cartilius 
Euplus in the temple in Ostia: the god reclines, naked and emasculate, 
beside the River Gallus, and is wearing a Phrygian cap fitted with solar 
rays (PL 1). 3C Indeed, the very earliest representation of this type, the 
bust of Attis attached to a dish in the Hildesheim treasure, now in the 
Antikenmuseum Berlin, where the god's Phrygian cap is decorated 
with stars, is generally dated to the late first century AD, if not even 
a little earlier. 37 References to Attis' cosmic role indeed usually take 
this form, of stars (and once the moon), or an eagle, represented on 
his Phrygian cap, or on his clothes. 38 The Naassene 'hymns' to Attis, 
which I have already mentioned, and which are generally dated to the 
second century AD, also clearly allude to this feature of the cosmology. 
At the close of the second fragment, for example, he is referred to as 
7toi|a,f|V A,£1)kcov ocoxpcov, 'shepherd of the shining stars'. 39 Elsewhere, 
the author provides the Naassene exegesis of the phrase, namely that 



36 Vat. Mus. MGP, inv. 10785. Found in a special deposit with two other items 
(probably for safe-keeping) in the portico on the S. side of the Campus. See Helbig 1 
1: 827 no. 1153 (E. Simon) = Vermaseren 1977: 61 with pi. 44 = CCCA 3: 123 no. 
394 = Vermaseren 1986 no. 312 = Rieger 2004, 282 cat. no. MMA 3; 138-41 with 
pis. 107— 108c (good commentary). 

3/ U. Gehrig, Hildesheimer Silberschatz am dem Antikenmuseum 2 (Berlin 1980) no. 14; 
Schwertheim 1974, 244f. no. 218b (late P-early P); CCCA 6: 21 no. 65 with pl.XIII 
(late Hellenistic); Vermaseren 1986 no. 345 (late P-P). 

38 Julian, Or. 5. 165b7f says that the Mother of the Gods fell in love with Attis and 
gave him a Phrygian cap adorned with stars (xov dcrcEpcoTOV JtTXov); this cap is then 
equated with the sky that surrounds us, which is said to cover Attis' head; consequently 
the River Gallus must 'really' be the Milky Way. Phe most familiar image of the Phry- 
gian cap associated with astral/cosmic symbols, in this case, the moon and a star, is the 
altar from Isola Sacra illustrated in Floriani Squarciapino 1962 pi. IV6 (not referred to in 
the text!; cf. however p. 71, List of figures); cf. Sfameni Gasparro 1981: 389; Lancellotti 
2000: 274 n. 205. Stars also decorate the Phrygian cap on some statuettes and coins: 
Vermaseren 1966, 51 with pi. XXXII. 1 and 4 (statuette, undated); Vermaseren 1986 
no. 368 (coin of Ancyra, second half IP); and they sometimes appear on his clothes 
(Vermaseren 1966, 33 n. 2 with pi. XIV2: coin of Cyzicus, second half IP; p. 51 
with pi. XXXII. 2, statuette from Asia Minor). For the eagle, cf. R. Turcan, L'aigle 
du pileus, in de Boer and Edridge 1978, 1287-92 with pi. CCLVIII = CCCA 3: 102 
no. 358 (05 4144). 

39 [Hippolytus], Ref. 5.9.9 Marcovich = Sanzi 2003, 287: Cybele no. 41. 



SYSTEMS OF BELIEF 39 

his human imperfection is resolved by the self-castration that enables 
him to become apa£v69r|A/i)c;, hermaphrodite, and thus re-assume his 
divine perfection as Kocwf| Kiiaiq, kouvcn; avBpomoq, a new creature, 
a new human-being — though we have no reason to suppose that this 
characteristically gnostic interpretation was also available within the 
cult (Ref 5.7. 13-15). 40 

However, I do not want at this point to get bogged down in details. 
The actual content of the narratives is surely the best introduction to 
the nature of these divinities and their relation to the cosmic order. 
At the end of the chapter, I go on to discuss the place of human beings 
in this world, and in the world beyond. 

a. The Egyptian Cults 

i. The Myth of his and Osiris 

Summing up his account of the myth of Isis and Osiris, Plutarch 
observes that it contains accounts of ocrcopicov mi 7toc0cov, of helpless- 
ness and suffering (De hide 20, 358f3). To that extent, they can be taken 
as paradigmatic. Although they were originally independent deities, 
Isis and Osiris are found in association from the third millennium BC; 
the earliest documentary allusions to them occur in the Pyramid Texts 
(Kees 1952). From then on, we find numerous allusions to the myth, 
but none of them are as complete as that provided by Plutarch in De 
hide et Osiride, esp. chaps. 12-19, 355d-358d. 41 

Plutarch's aim was to offer a capacious version that would at the 
same time try to eliminate the inconsistencies (cf. Froidefond 1988, 15). 
It is therefore not so much a literary version of a popular narrative 
as an intellectual compilation by a writer working under the political 
pressures of the day 42 As such, the text is by no means pellucid: beneath 



10 Cf. Sfameni Gasparro 1981, 384-86, 388£, and Lancellotti 2000, 278-81. Sfa- 
meni Gasparro, who argues very much as I do for the early development of a cosmic 
Attis, unfortunately relies on the older, Antonine, date for the Parabiago lanx (389 
n. 45), cf. n. 344 below. 

11 Plutarch died shordy after AD 120 (cf. K. Ziegler, s.v. Plutarchos, RE 21 [1951] 
640f.) and it is now generally agreed that De hide was one of his last works; Griffiths 
1970, 16f, dates it to c. 120. 

42 As Ziegler points out (ibid. 845f), Plutarch had been a pupil of the Middle Platonist 
Ammonius, who was an Egyptian, and had himself visited Alexandria. Richter 2001 
sees the work as a straightforward assertion of the primacy of Greek philosophy over 
Egyptian cult. This is too simple: on the keen interest of the later Stoics and Middle 
Platonism in 'oriental' religions, which were taken to contain traces of the supposedly 



40 CHAPTER TWO 

the intelligible surface narrative there lurk depths that confer upon 
the work a suggestivity entirely appropriate to a story about superhu- 
man beings. It also possesses the merit of providing a rationalisation 
of the myth-variants at the very moment when the oriental cults were 
arriving at their floruit. It is of course not my intention to privilege 
the information provided by intellectualisers of religion. 43 For lack of 
alternatives, however, we must use what we have, according each text 
its just value. In that regard, Plutarch's project is not so much didactic 
as ostensibly objectivising, in an attempt to render comprehensible, that 
is, in terms of early second-century Graeco-Roman philosophy, a myth 
that was ultimately Egyptian, though by now widely disseminated in 
the Graeco-Roman world. 44 The popularity of Isis at this time over the 
entire Mediterranean area must, I think, have meant that his re-work- 
ing of the myth was greeted not as something bizarre but as desirable 
and needful. It was accepted by contemporaries precisely inasmuch as 
it could be seen as a guide for study, and not because it was in any 
way an official or canonical version. And that is the spirit in which we 
too should read it. 45 

The narrative, briefly told by Plutarch, is roughly as follows: 46 There 
were two pairs of divinities, all of them children of Rhea-Nut (the 
celestial great mother), who symbolise the opposing principles of Good 
and Evil. The first couple was composed of Osiris and his sister-wife 



authentic religion of primitive man, see G.R. Boys-Stones, Post-hellenistic Philosophy: A 
Study of its Development from the Stoics to Origen (Oxford 2001) 3-59; 99-122. 

43 I am thinking here of a point made by Hopkins 1999, 4: "In the Roman world 
stories, not analysis, were the stuff of religious persuasion". 

44 On the other hand, Griffiths also claims that Plutarch's speculations are foreign 
to the realities of the contemporary cult of Isis in Greece: "Only a very cultured elite, 
one can imagine, would have thought like this, and they would have been nurtured 
in Neo-Platonism before being converted to Isis and Osiris" (1970, 74). The reasons 
adduced relate solely to items of factual information, but there is nothing to hinder us 
from thinking that Plutarch's 'theological' concerns were shared by leaders of the cult, 
for example, the Isiac priestess Klea, to whom the treatise is dedicated (35, 364e). She 
is usually identified (admittedly only on account of the homonymity) with the Flavia 
Klea who appears on a couple of contemporary inscriptions from Delphi (Griffiths 
1970, 17; Froidefond 1988, 19ff). 

45 See however the remarks of Richter 2001. 

46 I mainly follow Plutarch here, incorporating relevant details from Diodorus 
Siculus, together with some explanatory comments. One should constantly bear in 
mind Griffiths' very full commentary, already mentioned (Griffiths 1970), as well as its 
predecessor, Hopfner, 1940—41, the first volume of which is devoted to a commentary 
on the myth, with the Greek text of chap. 12—20 and a translation into German, while 
the second is devoted to the rest of work, albeit without a Greek text. Froidefond 1988 
is also useful. 



SYSTEMS OF BELIEF 4 1 

Isis, who fell in love even before they were born and secretly mated in 
the darkness of the womb. This must be an allusion to heterosexual 
endogamy inserted later as a mythical legitimation of brother-sister 
marriage in the Pharaonic world, as Diodorus points out (1.27. 1). 4 ' 
The other pair consisted of the evil Seth, translated into Greek as 
Typhon, and his sister-wife Nephthys, who are also the siblings of the 
good pair — four siblings, therefore, all of whom turn out to have sexual 
relations. 48 Osiris, the lord of all things, reigned wisely in Egypt. He 
taught the Egyptians agriculture, gave them laws and instructed them 
in the worship of the gods. He then decided that he ought to set out on 
a mission to civilise the rest of the world by means of friendly persua- 
sion. His honeyed words, accompanied by music and song, enabled him 
to manage the task more or less without violence, so that the Greeks 
identified him with Dionysus. 49 Isis is in turn identified with Demeter 
(Diod. Sic. 1.25.1; 96.5), so that Plutarch's treatise is in good part a 
justification of the kinship between the Egyptian and the Eleusinian 
mysteries. 50 

During his absence, Isis ruled as regent in his stead (cf. Diod. Sic. 
1.17.3). On his return however Seth organised a conspiracy in which 
he was joined by seventy-two others. In order to get rid of his brother, 
he devised the following ruse. Having secretly measured Osiris up, he 
had the most beautiful sarcophagus imaginable constructed. He then 
invited his fellow-conspirators to a feast on 1 7th Athyr (1 3th November). 
Osiris too was to be there. As the feast took its merry way, Seth had 
his creation brought in. The guests fell silent with wonder, and Seth 
took advantage of the situation to offer it as a gift to anyone present 
who fitted it exactly. One after another the guests tried it out until it 



" It is not certain that the practice existed during the Pharaonic period outside 
the limits of the royal house: J. Cerny, Consangineous Marriages in Pharaonic Egypt, 
JEA 40 (1954) 23—29. In the Graeco-Roman period it became a general practice: H.I. 
Bell, Brother and Sister Marriage in Graeco-Roman Egypt, RIDA 2 (1949) 83-92; 
Griffiths 1970, 308f.; S.L. Ager, Familiarity Breeds: Incest and the Ptolemaic Dynasty, 
JHS 125 (2005) 1-34. 

48 On Nephthys, the youngest of the four children of Geb and Nut, see briefly 
Bonnet 1952, s.v. 

49 Cf. Hdt. 2.42; 144; 156, though he says that the identification was due to the 
Egyptians; Diod. Sic. 1.1 4— 18 ascribes Osiris a role in Dionysus' adventures. 

50 Griffiths notes (1970, 309) that the passage devoted to Osiris' civilising mission is 
paralleled in Diod. Sic. 1.14., cf. also Froidefond 1988: 68ff The common source may 
have been Hecataeus of Abdera. The material is presented in his own distinctively 
unreliable manner by Merkelbach 1995, 37—55. 



42 CHAPTER TWO 

came to the turn of Osiris, as naive as he was good. All unsuspecting, 
the king stretched himself out in the sarcophagus; at that moment the 
conspirators sprang the trap, clapped on the lid and sealed it with lead. 
Thus perished Osiris, a miserable end. To avoid possible trouble and the 
better to attain his further ends, Seth shoved the sarcophagus into the 
Nile. Bobbing on the waves, it was carried down to the river's Tanaitic 
mouth in the Delta, which as a result is named 'Accursed'. The box 
then floated out to sea until it was trapped beneath an 'erica-tree' in 
the Phoenician city of Byblos, far up the Levantine coast. 51 

Bearing in mind that this myth was standardised, at least to an extent, 
already in the third millennium, it has been suggested that this odd 
ending might be an allusion to the relations between Byblos and Egypt 
at that time, encouraged as they were by the natural currents of the 
region, which flow up the Syrian coast in an anti-clockwise direction. 32 
No Egyptian source however mentions the role of Byblos in connection 
with the discovery of the sarcophagus — Plutarch is actually the earliest 
evidence for it. There could be various reasons for the box's turning 
up at Byblos, for example to account for the celebration of Egyptian 
cult there (attested for Osiris at least since the New Kingdom, and 
for Isis from the seventh century BC), or perhaps the assimilation of 
Osiris and Adonis. 53 

When Isis heard the news, she cut off a lock of hair and dressed 
herself in mourning. Feverishly but in vain she hunted for her husband: 
no one knew his whereabouts, save some children who showed her the 
direction that had been taken by the sarcophagus. That is the reason why 
the Egyptians believe that children have prophetic powers, especially 
when they shout their heads off while playing in sacred areas — schools 
were often attached to temples (cf. Griffiths 1970, 315). In the mean- 



51 epeiKri means 'heather', but the implausibility of this has led to various alternative 
suggestions, including 'cedar' and 'tamarisk', on the assumption that it is a mistransla- 
tion of some Egyptian or Semitic word. 

52 Cf. P. Montet, Byblos et I'Egypte (Paris 1928) 271, citing documentation from the 
IV Dynasty onwards; J. Pirenne, Histoire de la civilisation de I'Egypte ancienne (Paris 1961) 
54ff ; more attention to the archaeology in I.E.S. Edwards, The early Dynastic period 
in Egypt, CAH 2 1.2 (1985) 45ff M. Silver, The Mythical Conflict between Osiris and 
Seth, and Egypt's Trade with Byblos during the Old Kingdom, in idem (ed.), Ancient 
Economy in Mythology: East and West (Savage, MD 1991) 193-215 defends the story's 
historicity, ignoring the competing traditions. In his view, the key is not funerary 
symbolism but trading-links: each of the gods represents a vector in the commercial 
relations between Egypt and Phoenicia. 

53 On Osiris and Adonis, see the careful discussion by Lightfoot 2003, 305-28. 



SYSTEMS OF BELIEF 43 

time, Isis learned that her beloved Osiris had enjoyed the favours of 
Nephthys, mistaking his sister/ sister-in-law for his sister /wife. She also 
learned that the child that resulted from this case of mistaken identity 
had been abandoned by his mother immediately after the birth, for 
fear of Seth/Typhon. All this is mentioned in a digression that has no 
connection with Plutarch's narrative at this point, which has led some to 
suppose that Seth's act, unmotivated except by his general wickedness, 
was really in revenge for Nephthys' adultery with Osiris. 34 In this con- 
nection, however, we cannot ignore the Songs of Isis and Nephthys, scraps 
of which are found in several different papyri: they refer unmistakably 
to the adultery and imply that it was actually an affair. In other words, 
Plutarch, moralist that he is, has deliberately suppressed it; Diodorus, 
earlier, implies knowledge of the story in stressing that Isis swore never 
to have carnal relations with another man after the death of Osiris: 
utiSevoc; dvSpoc;, with no other man — but says nothing of capricious 
gods (1.22.1, repeated at 27.1). At any rate, after a great deal of effort 
and with the aid of some dogs, Isis eventually tracks down her stepson 
and twice-over nephew, the divine Anubis, who thereupon becomes 
her most faithful and discreet guardian and accompanies her on her 
subsequent adventures (Diod. Sic. 1.87.2-3). 

The children's information proved to be correct, and so Isis set out for 
Byblos. Once there, she managed to find Osiris thanks to her knowledge 
of magic. The coffin had become trapped inside the base of the 'erica- 
tree', which had shot up with extraordinary rapidity; this miraculous 
growth had attracted the attention of the king, who arranged for the 
lower section of the trunk, including the part enclosing the coffin, to 
be installed in his palace as a pillar to support the roof. Isis managed 
to gain entry to the palace, and was asked by the queen to wet-nurse 
her baby, the second of two boys. She did so, using for the task not her 
nipple, but a finger. 35 Deciding to render the young prince immortal 



54 So Hopfner 1940-1, 1: 46, rejected by Griffiths 1970, 316f. Both authors refer to 
the papyrological evidence for Nephthys pleading with Horus to open the door so that 
she can come in to his father; or 'taking care of her brother'; that Osiris was in love 
with (Sene)nephthys; and for Isis complaining to her father Thoth that Nephthys has 
slept with Osiris — all of them aspects of this edifying divine cacophony. 

55 It has been claimed that this is an allusion to a rite of adoption, of a type attested 
in Ethiopia in the 1860s; but, as Griffiths 1970, 327 points out, there is no ancient 
evidence for it. The gesture of Harpocrates, in putting his finger to his mouth, has 
no connection with this mythic detail. If anything, it may have been a sign requesting 
adoption. 



44 CHAPTER TWO 

by off burning his mortal parts, she put him into the fire, and, turn- 
ing herself into a swallow — both in Egypt and in Greece an emblem 
of mourning (Griffiths 1970, 328f.) — flew, twittering with grief, round 
and round the pillar that concealed the remains of Osiris. This process 
evidently took some time. One night the queen saw her, and screamed 
in alarm, thus depriving the child of immortality and forcing Isis to 
reveal herself. 56 At that, the goddess demanded the coffin, cut it out of 
the tree-pillar and threw herself over it with cries of such penetrating 
grief that the baby prince was unable to bear it and expired. 

Distracted by sorrow, Isis set sail for Egypt with the sarcophagus. In 
the Delta marsh, at Buto, she met her son Horus/Harsiesis. This Horus 
is neither the one Plutarch mentions at the beginning as the brother 
of Osiris and Isis, nor the one who is going to appear in the form of 
Harpocrates ('Horus-the-child'). This fact, emphasised by all the com- 
mentators, is highly relevant to Pharaonic theology, for it alludes to the 
new Pharaoh; 57 but it does cause problems of interpretation in con- 
nection with the combat with Seth and Horus' extraordinary violence 
against Isis, as we shall see. 

Isis resolved to hide the sarcophagus, but unfortunately Seth stumbled 
upon it one night while out hunting. In a fury he removed the body and 
ripped it into fourteen parts, which he proceeded to scatter about over 
the entire country 58 Isis then began a sad tour up the Nile, protected 
by the crocodiles, in order to locate the pieces of her dismembered 
brother-husband. One by one she discovered them and gave them 
burial, either, as one version has it, where she found them, or by having 
statues of Osiris made out of aromatic wax and establishing a defunc- 
tive cult at each site, in order to prevent Seth from discovering which 
tomb was the real one, and to ensure that the cult became as widely 
diffused as possible. 39 



56 The relation between this part of the myth and the Homeric Hymn to Demeter esp. 
169-255 has been noted by S. Herrmann, Isis in Byblos, %AS 82 (1957) 48-55. 

57 The deceased Pharaoh is Osiris, the new one is identified with Horus, cf. Griffiths 
1970, 337f.; Froidefond 1988, 269 n. 7; Merkelbach 1995, 87-93. Griffiths however 
rightly observes, "Harpocrates ... is not easy to distinguish from Harsiesis save in that 
he is consistently depicted as a child" (p. 353). 

58 The number 1 3 might have to do with the number of nomes, the territorial and 
administrative units of Egypt. But different versions offer different numbers: Diodorus 
1.21.2, for example gives 26; in the Osiris temple at Dendera, during the Ptolemaic 
period, we find 16 in one text, 14 in another; at Edfu, we find the total of 42; cf. 
Griffiths 1970, 338. 

,9 Diod. Sic. 1.21.5—11; Strabo 17.1.23, 803C; cf. Seneca, de superstitione ap. Serv. 
ad Am. 6.154. 



SYSTEMS OF BELIEF 45 

The only part she could not find was the penis (Diod. Sic. 1.2 1.5), 
because Seth had tossed it into the Nile and it had been gobbled up 
by some fish. 60 So she made a replica, and set it up for worship, which 
continued down to Plutarch's day. At the same time, the tradition that 
Harpocrates was conceived after Osiris' death shows that the severed 
penis, for all that it had been eaten, continued in some sense to exist 
and retain its virility (cf. Casadio 2003, 256-58). The injunction 
against eating certain kinds of long, thin fish might be related to their 
resemblance to a phallus, but the idea is not very persuasive, since the 
prohibition only applied to specific fish in certain areas of Egypt. On 
the other hand, the fact that the only part of Osiris that almighty Isis 
failed to recover is the naughty tool seems, in the extant accounts, to 
have nothing to do with the adultery. To be sure, in ordinary life in 
Egypt, adultery seems to have been severely sanctioned: according to 
a story in EWestcar, an adulterous woman was burned and her ashes 
thrown into the Nile (Griffiths 1970, 317). It is however far from clear 
whether this fact (the story applies to the woman, not the man) is rel- 
evant to the interpretation of a foundation-myth of the type we are 
dealing with. Since the ancient texts do not make an explicit connec- 
tion between the adultery of Osiris and Nephthys and the loss of the 
penis, we cannot know whether the followers of Isis made one. 61 But an 
irreverent joke of this kind during the singing of the songs that recalled 
the famous adultery and the fate of the tool of divine bliss would not 
be implausible. A quick mind prefers humour to its moralising opposite: 
the fate of Osiris' penis might have been found amusing since everyone 
knew that he had been unfaithful to Isis, for all that the connection is 
not made explicit. Whatever the truth here, the primary connotation 
of the penis acquiring a life of its own and being able to engender 
Harpocrates is its miraculous power of fecundation, that far surpasses 
ordinary human understanding. In this context, Osiris is 'the mummy 
with the long phallus' (Griffiths 1970, 343f.). 62 



60 According to Plutarch, this is the reason why the lepidotos, the phagros and the 
oxyrhynchos were not eaten (De hide 18, 358b). 

1)1 Several ancient texts would legitimate the term 'Isiacs' to denote the followers of 
Isis, e.g. Valerius Maximus 7.3.8; Suetonius, Dom. 1.2; ILS 1260 (Fabia Aconia Paulina, 
wife of Vettius Agorius Praetextatus); 4369, 641 9f; 6420b. [Alvar here uses the substan- 
tive 'isiacos' (plur.), which is plausible in Spanish, but appears uncouth in English. I 
have continued to use the usual periphrasis 'followers of Isis'. Tr.] 

62 Note also Diod. Sic 1.22.6-7. Burton 1972, 96f observes that in Egypt there is little 
narrative trace of the loss of Osiris' phallus, and no marked cult of it except at Mendes. 



46 CHAPTER TWO 

Thanks to her untiring efforts, Isis succeeded in gathering together the 
dismembered remains of her brother-husband. This sorrowful progress 
came to be the focus of the inventio Osiridis, the ritual celebrating the 
miraculous discovery of the fragments of Osiris' body with the aid of 
the faithful Anubis. Quest and discovery were the essential preliminar- 
ies to the recomposition of the corpse, which was thus able to return 
to life, or rather to bestow new life, in the sense that Isis could have 
intercourse with the inert corpse and thus become pregnant. 63 The 
symbolic meaning here is clear enough: salvation in the cult of Isis is 
only possible on the basis of the individual's voluntary search for truth, 
and — more important still — of the goddess' divine aid, which, thanks 
to her magical and medicinal arts, makes resurrection and immortality 
possible (Diod. Sic. 1.25.2-7). Obviously this is not a resurrection that 
reverses the rules of the real world. 64 Once dead, Osiris could not be 
restored to his social existence, however greatly his kin might wish it, 
since that has gone for good (this is one of the difficulties posed for the 
imaginaire by any attempt somehow to preserve the dead in a stronger 
form than mere memory). One way of reducing the problem is to create 
a fictional world, of the living dead able to communicate through the 
mind; yet, although at first sight such a solution appears to offer relief, 
ultimately it only makes the already burdensome world of mortals still 
more fraught. The fantasy of meeting again in an agreeable world 
beyond this one does provide a sort of consolation, but at the same 
time it crams this one full of ghosts. 

Religion loves to play with inconsistencies. Although Osiris has 
descended to the one fitting place, the Underworld, myth is allowed 
to contradict itself. For Osiris returns from there in order to complete 



63 Cf. Frankfort 1948, 40 and 356 fig. 18; Griffiths 1970, 343; at 353 he cites 
P.Louvre 3079, where the goddess says, "I have played the part of a man though I am 
a woman, in order to wake thy (Osiris') name since thy divine seed was in my body". 
This seems to mean that in bringing about Osiris' revival she played both the male 
and female part in the sexual act. It is even possible that the posthumous conception 
of Horus is alluded to in the Pyramid Texts (632a-d): "Thy sister Isis has come to 
thee, joyous through love of thee. She places for thee thy phallus on her vulva. Thy 
seed comes forth into her, so that she is equipped as Sothis". Griffiths however points 
out that the allusion to Sothis = Sirius proves the astral character of the text, Osiris 
being Orion (1970, 353 n. 6); cf. G. Glerc, Isis-Sothis dans le monde romain, in de 
Boer and Edridge 1978, 1: 247-81. 

1,1 Diodorus however explicitly says that Isis managed to bring her son Horus (= 
Harsiesis) back to life and made him immortal (1.25.6 with Burton 1972, 109; cf. 
POxy 1380 1.246E). 



SYSTEMS OF BELIEF 47 

a task that has hitherto scarcely been mentioned, the completion of 
the education of his son Horus/Harsiesis, whom Isis met in the Delta 
on her return from Byblos. 65 The dead king does not seem to able to 
rest so long as justice does not prevail on earth: Horus appears not so 
much the avenger of his father as his helper or supporter. 66 At any rate, 
Osiris comes to Horus to train him to become a fine warrior; since it is 
not possible fully to restore the social relations of the dead, Isis is not 
mentioned. 6 ' However Horus has now clearly taken over responsibility 
for the family. He is the young man who guarantees continuity (as rep- 
resented by the existence of three distinct Horuses), whose responsibil- 
ity is to enable his parents to enjoy peace by confronting Seth. 68 This 
evident emphasis on the values of the family, as a structure where the 
basic rules of life-sharing and continuity through time are lived out, 
gives us a good insight into the values that underlie the Roman cult of 
Isis. At the same time, values such as these are not limited to the family 
and can become a political programme: Horus is at the same time the 
living Pharaoh, whose right to the throne has been challenged by Seth, 
acting to all appearances like a typical uncle. The Ennead, the council 
of nine gods, manages to resolve the complex conflict, now involving 
the issue of order at three different levels, of the dead, of the dynasty, 
and of the cosmos, as well as that of good versus evil. We might add 
that the social, ethical and political implications of such questions were 
of the greatest interest in the High Empire, when the myth was being 
reformulated in this manner. 

Overcoming evil requires violence — lots of it. Day after day Horus 
and Seth, in the form of hippopotami, fight titanic battles, causing a 
tremendous amount of damage. The heavens tremble, and the gods in 
alarm decide, in one version of the myth, to separate the contenders 
by creating a no-man's-land between them. In the end Horus manages 
to overcome Seth and put him out of action. 69 But then Isis feels pity 



65 That the Horus of chap. 19 is Harsiesis is made clear by the mention right at 
the end of Isis' giving birth to Harpocrates (358d9-el). 

66 So rightly Griffiths 1970, 344f. on De hide 19, 358bl0f. 

67 She is however found helping Horus towards victory in Diod. Sic. 1.88.6, but the 
allusion is too brief to justify withdrawing the claim in the text. 

68 Even the Egyptian sources sometimes get muddled up over the three Horuses, 
as is clear from the 'Contendings of Horus and Seth' in PChester Beatty no. 1 (Middle 
Kingdom): Lefebvre 1949, 195 n. 72. 

69 The most complete narrative is to be found in the 'Contendings' (see previous 
note) in Gardiner 1931, of which there is an expurgated French translation in Lefebvre 's 
collection. ESallier IV however also contains a good deal of information. 



48 CHAPTER TWO 

on seeing her brother Seth bound in chains, and sets him free. Such 
female weakness, even on the part of a goddess, his own mother, infu- 
riates Horus, who has been badly knocked up in the struggle (as we 
learn from the Egyptian texts; Plutarch says nothing about it), and he 
attacks her, in one version, including Plutarch's, tearing off her crown, 
in another, lopping off her head with his 1 61b chopper. In the versions 
in which she is decapitated, Thoth/Hermes provides her with the head 
of a cow (i.e. Hathor, who was the mother of yet another form of 
Horus), which is the basis of her usual identification with Io, daughter 
of Inachus, in Latin poetry 70 

Horus' fury, which has sometimes been interpreted as a response to his 
own probable illegitimacy, was prompted rather by Isis' soft-heartedness 
towards Seth during the battle: according to one version, Horus was 
infuriated by the fact that Isis, having stuck a barb into Seth, took it 
out again when he begged her to do so (Griffiths 1970, 349f). Seeking 
to explain their enmity, Plutarch claims indeed that Seth prosecuted 
Horus before the gods on a charge of being illegitimate, though there 
is no support for this in Egyptian sources (Griffiths 1970, 351f). The 
motif of fury allows Plutarch to suppress any allusion to the homosexual 
relation between Seth and Horus," the castration and dismemberment 
of Seth, the amputation of Horus' hands and his seduction by Isis, all 
of them constituent elements of the myth in its local variants, that run 
through the various possible patterns of sexual relations which are the 
gods' prerogative even as they destabilise the order of the cosmos. 72 

However I do not think it has been sufficiently stressed that Horus' 
disapproval might be related to the danger implicit in his mother's new 
pregnancy. In political terms, it is his succession to the throne that is in 
jeopardy, since Isis is carrying in her womb a potential new Pharaoh; 
socially, however, it is doubtful, as some Egyptian sources put it, whether 
the wife could have become pregnant by her husband once he was 



70 See the references in T. Gantz, Early Greek Myth (Baltimore and London 1993) 
198-204. 

71 See PChester Beatty II 1—12, recorded also in a text of the XII Dynasty: Lefebvre 
1949, 180: see further Griffiths I960; W. Barta, Zur Reziprozitat der Homosexuellen 
zwischen Horus und Seth, Gbttinger Miszellen 129 (1992) 33-38: Montserrat 1996, 142f. 
On Plutarch's intentions in suppressing 'barbaric' details, cf. briefly Richter 2001, 
207. 

72 The struggle between Horus and Seth is primarily for sovereignty over Egypt; 
in some versions, once defeated, Seth is treated honourably by Horus, in others he is 
expelled and/or mutilated (Griffith 1970, 349). 



SYSTEMS OF BELIEF 49 

dead.' 3 If her behaviour had been improper, Horus, as king, would be 
obliged to punish his mother's indiscretion, by bringing about her death 
or the loss of her crown, which are symbolic equivalents. The disagree- 
ments between the traditions that claim that Isis is supreme and those 
that promote Horus reproduce the theological and territorial conflicts 
that are so characteristic of Egyptian history, and naturally re-appear, 
complete with their irreconcilable features, in the mythical discourse that 
was re-elaborated in the Hellenistic-Roman period. Religion however 
can project these contradictions onto a more profoundly symbolic world 
beyond, and allegorical readings extracted from them so as to keep the 
faithful lulled in the opacities of the incomprehensible, subject to the 
order postulated by the unfathomable mysteries of religion. 

Though he himself lost both his testicles in the battle, 74 Seth tore 
out one of Horus' eyes; once out, it set off to voyage up through the 
heavens, and became the sun's disc. This is the key to understanding 
one of the meanings of the myth. Horus is the sun itself, whose triumph 
overcomes chaos and creates the order of the annual cycle that allows 
human-beings to know in advance what is to occur at each season of 
the year. Isis is another astral symbol, the moon. 75 The myth's remaining 
personifications represent a succession of other planets or explanations 
of the succession of the seasons. There can be no doubt that this is 
a mythical account of the cycle of the year and of the relationships 
between certain stars. That is, cosmic order returns after the chaotic 
conflict with the triumph of Horus, who represents the living Pharaoh, 
the son of Osiris; the latter is the dead Pharaoh and the god of the 
underworld, thus linking him with a specific form of fertility. 



' 3 Claimed both by P.Chester Beatty I and ESallier IV 

74 Gardiner 1931, 20 no. 5 and 21 no. 1, cf. Griffiths 1960, 29. In this version, 
Seth tears out both of Horus' eyes, which turn into bulbs from which lotus-plants 
emerge. 

" This is explicitly stated by a controversial passage of Diodorus, 1.11.1, where he 
says that Isis is the moon and Osiris the sun. Etymologically this is indefensible, and the 
doubt has then spread to the claim that Osiris was a solar divinity. I think that those 
who reject this evidence are right, but that Diodorus' 'nonsense' can be minimised by 
assuming that the heavenly bodies he refers to, the sun and the moon, are not the dead 
Pharaoh Osiris and Isis, but the former's living son, Horus, and Isis. The complexity 
of the family-relationships in the myth may have contributed to Diodorus' (or rather 
his source's) 'mistake'. In my view this is the least violent means of resolving the dif- 
ficulties presented by the passage. 



50 CHAPTER TWO 

Osiris' role as a divinity connected with growing things associates him 
in particular with grain.' 6 At the level of explicit narrative, it is he who 
introduces agriculture to Egypt; but the myth also replicates the annual 
vegetative cycle by representing Osiris as a god who gives life after a 
transitional period: he lives under the earth for a period of time and 
there germinates the seed that will give new life. 77 The myth's proleptic 
rehearsal of the sequence required for successful agricultural production 
renders the mystery of life intelligible. On the other hand, one may 
also hold water to be the ultimate cause of the regeneration of nature; 
Osiris plays a role here too, being present in the liturgy through the 
hydria, which contains the token Nile-water, identified with Osiris. 78 

The myth thus reproduces the biological cycle, whose basis is the 
family. The family is indeed the elementary form of all social relations. 
If it holds together, any problem, however serious, can be surmounted; 
conversely, any act that violates that bond may provoke incalculable 
disasters. This is the second basic level of reading myth: as an expla- 



"' Cf. Athenagoras, Legatio/Apologia 22.9: 'Osiris is the sowing of the wheat', 
naturally emphasised by Frazer 1914, 2: 30-48, and esp. 96-114; Hopfher 1940-41, 
2: 250-54; Frankfort 1948, 185-90; J.G. Griffiths, The Origins of Osiris and his Cult, 
Numen Supplement 40 (Leyden 1980) 163-70; Hani 1976, 155-58. However Griffiths 
has solid arguments to show that the agrarian character of Osiris cannot be of very 
great antiquity, since he does not seem to be associated with barley until the Middle 
Kingdom (cf. the Ramesseum Dramatic Papyrus 29-33 [ed. K. Sethe, Dramatische Texte zu 
altdgyptischen Mysterienspielen, Leipzig 1928]). Indeed, it has been said that vegetation is 
a Nebensache as far as he is concerned (Mettinger 2001, 169f). The original reason for 
the linkage seems to be the Egyptian calendar's correlation with the annual flooding 
of the Nile, which is to be connected with Osiris' resurrection: cf. Plutarch, De hide 
39, 372c with Bonneau 1964, 245-48. The gradual slippage of the calendar meant 
that the festival ended up coinciding with the Nile subsidence, the moment when 
things began to grow, so that the resurrection came to be linked to the growing-cycle 
(see also Chap. 4. 4. a below). Isis is obviously also associated with plant-growth, since 
some traditions claim that her tears were the cause of the Nile flood. A more direct 
link to the agrarian sphere is however her identification with the serpent Renenutet, 
better known as Thermouthis: PJumilhac 23.13 says that Isis turned into an uraeus 
(cobra), her sacred animal (Aelian, de nat. anim. 10.31; Apuleius, Met. 11.3; cf. Ovid, 
Am. 2.13.13; Met. 9. 694 with Bomer ad loc; Juvenal, Sat. 6.538 with Friedlander ad 
loc). Isis-Thermouthis is not attested before the Ptolemaic period, though she may 
have been created as early as the New Kingdom, cf. G. Deschenes, Isis Thermouthis: 
a propos d'une statuette dans la collection du prof. M.J. Vermaseren, in de Boer and 
Edridge 1978, 1: 305-15; Malaise 1985, 125-55. A selection of images in Mostra hide 
232f. nos. IV228-30. 

" Mettinger reproduces a bas-relief from the roof-temple at Philae, which shows 
the corn-ears sprouting out of his corpse, and being watered by a farmer (2001, 171 
fig. 6.1). 

,s Plutarch, De hide 32, 363d; Griffiths 1970, 420f collects various Egyptian texts 
that identify Osiris with the Nile-god; see further Chap. 4.4.b.i below. 



SYSTEMS OF BELIEF 5 1 

nation of how the world functions; or rather, as a justification of the 
established order of things. Because myths are stories that offer examples 
of normative social behaviour, they contain an implicit system of values 
or standards. Although Horus and Osiris represent the continuity of 
the power of the Pharaoh, the reading of the myth in the Principate 
focused more on relationships of inter-personal power (though politi- 
cal relationships are not entirely absent). Her repeated flirtations with 
the imperial cult notwithstanding, Isis' dominant position shifts the 
emphasis onto the domestic side of things. The main image is of 
Isis sustaining the order of this world, controlling the realm of social 
relations where her power is particularly manifest. It is because Isis is 
assigned this function in myth that both Plutarch and Diodorus stress 
her virtues as a wife. 

The well-known self-predication of Isis from the temple of Isis and 
Serapis at Cyrene, a metrical text dating from AD 103, illustrates how 
her adherents managed to integrate all the praiseworthy features that 
I have mentioned so far: 

I, Isis, am sole ruler of time, inspector of the limits of the sea and the 
land, and, with sceptre in hand, their sole inspector. All name me supreme 
Goddess, the greatest of all the gods in heaven. For I myself have discov- 
ered everything, all is my work: the writing on the seals shows it clearly, 
revealing to all the inventions I have vouchsafed to mortals, and the fruits 
of the soil. I have fortified cities with reverend walls, and to mortals I 
have shown how to understand (such skills) clearly. Without me nothing 
has come into existence, and the stars do not hold their courses without 
first receiving my instructions . . , 79 

Finally, when she has re-ordered the cosmos and regulated social rela- 
tionships in the world, all that is missing is control over life beyond the 
grave. The transition between the sensible world and the Underworld 
is in the hands of Anubis, the bastard son of Osiris, who, in his role as 
psychopomp, leads mortal souls into the presence of his father, the god 
of the realm of everlasting gloom. A text not often cited in this context 
clearly distinguishes between two levels of the gods' activity in a manner 
suggesting that they were conceptually distinct. In his treatise on the 
interpretation of dreams, Artemidorus distinguishes between the physi- 
cal teaching of the Egyptian gods and the mythical (Oneirocrit. 2. 39). 



79 SEG 9: 192 = Totti no. 4 = RICIS 701/0103, cf. E. Des Places, La religion grecque 
(Paris 1969) 164f. 



52 CHAPTER TWO 

Unfortunately this is not very specific, but he implies that the cult 
celebrated grief, and, as a result, the appearance in a dream of Isis, 
Serapis, Anubis or Harpocrates foretells threats and danger. We may 
suggest that, since when we are awake they instead afford salvation 
from evils, Artemidorus' 'physical teaching' is the opposite of pain and 
the suffering it entails. 80 

In any case, we have seen how a solution is provided in mythical 
terms for the troubles that reality and imagination bring upon human- 
beings. Evil is given a basis and a certain number of motives. Once it 
gains a hold, it subverts all forms of order, but the ancient notion of 
Ma'at (cosmic harmony) can be restored thanks to the sufferings of the 
divine family that for human communities symbolises perfection not 
merely while they are here on earth but also, especially in the Roman 
period, in fearful eternity. 

ii. Serapis 81 

Alexander's conquest of Egypt initiated a process of culture-contact 
between Greeks and Egyptians, spanning a continuum between mere co- 
existence (the incidence of bilingualism was quite limited) and habitual 
or regular assimilation of cultural practices. At the socio-economic 
level, the co-existence turned into what was clearly an oppression, 
from which the sole escape was a form of local oligarchy. This created 
serious political problems, which were palliated by a variety of means, 
including the encouragement of a hybrid culture focused on the new 
city of Alexandria. As a new foundation, Alexandria stood in need of 
a tutelary deity. It was desirable that such a deity should incorporate 
elements of Egyptian tradition in order to fit in with the local gods, 
but also possess Greek traits so as not to appear odd to the new colo- 
nists; the aim was that they should accept him as their own deity. The 
new god had to be as acceptable in Alexandria as in the surrounding 



80 A little further on, however, he says much the same about Cybele. On divine 
intervention in the structure of dreams and its relation to salvation in the mysteries, 
see Hidalgo 1992. 

81 The proper form of the god's name has been the subject of much controversy; a 
whole range of different etymologies has been proposed starting with the speculations 
of ancient authors, cf. Dunand 1973, 1: 5 Iff; G. Mussies, Some Notes on the Name 
of Sarapis, in de Boer and Edridge 1978, 821—32. I use the later form Serapis, since 
it is the one most commonly found in Latin epigraphy; although Sarapis is usual in 
Greek, Serapis also occurs there. 



SYSTEMS OF BELIEF 53 

territory (chord). 82 All this explains the new regime's decision to act 
rapidly, for there can be little doubt that it was Ptolemy I Soter who 
first established the cult. The King is supposed to have received the 
command in a dream (Plutarch, De hide 28, 362a). The later tradition 
produced a fine array of etymologies to account for the name. It was 
claimed that Ptolemy took delivery of a statue of Pluto from the city 
of Sinope, since this was the god who may have appeared to him in a 
dream and bade him seek him out. 83 After the transfer to Alexandria, 
the statue was given the name Serapis, supposedly the Egyptian name 
for Pluto. This identification is repeated in all the contemporary sources 
that mention the new god (Dunand 1973, 1: 46f). 

There were a number of quite different versions of Serapis' ori- 
gins. 84 Few nowadays credit the story that the friends of Alexander 
had consulted the god in his temple in Babylonia when the King was 
dying 85 It has been suggested that this was some kind of oracular Baal 
subsequently identified with Serapis. By this account, Serapis might 
have been taken to Sinope from Babylonia, and thence to Alexandria. 
Most scholars however have tried to solve the problem by distinguishing 
between the introduction of the god's statue at the end of the fourth 
century BC and the 'creation' of a new cult, most likely in the reign 
of Ptolemy II. 86 It seems clear that there were at least two, perhaps 
several, quite different accounts which have become indissolubly fused, 
so that there can be no final clarity. 

There is little sense to be made of the claim that a statue of Pluto 
was brought from Sinope to Alexandria in the late fourth century BC, 



82 Vidman 1970, 45 (this section provides an excellent account of the historical 
context of the creation of Serapis). The view that the cult was created with a view to 
bringing Greeks and Egyptians together was argued already by P. Jouguet, La politique 
interieure du ler Ptolemee, BIFAO 30 (1930) 513-36. Whether it succeeded is of course 
another question. It has been argued, for example, that the Serapeum mentioned by 
Zenon in his letter to the dioiketes Apollonios (SB III 6713, 257 BC) is one that was to 
be built in the Greek quarter of Memphis, to serve the Greek population exclusively: 
K.J. Rigsby, Founding a Serapeum, GRBS 42 (2001) 1 1 7-24. 

83 Clement of Alexandria, Protr. 4. 48.2 Stahelin, p. 74. 5f. Marcovich = Sanzi 2003, 
80f. no. 7, says that it was Ptolemy II Philadelphus, not Soter; Tacitus, Hist. 4. 83f. seems 
already to reflect sources that claimed the cult was founded by Ptolemy III. Eusebius' 
account is muddled. The convoluted state of the tradition makes it impossible to be 
absolutely confident; for a recent account, see Scheer 2000, 260-66. 

81 Fraser 1972, 246—76 offers in my view the best, most balanced synthesis of the 
whole tangled topic; also Stambaugh 1972. On the iconography, Hornbostel 1973 is 
dreadfully long-winded; Clerc and Leclant 1994 much to be preferred. 

85 Plutarch, Alex. 76.9, 706f; Arrian, Anab. 7.26.2. 

86 E.g. Stambaugh 1972, 6-13; Dunand 1973, 1: 48. 



54 CHAPTER TWO 

and that thirty years later a new cult developed around it. It is surely 
more plausible to think that various syncretistic pressures, both native- 
Egyptian and Graeco-Egyptian, put their stamp on the new god, 
who, perhaps originally conceived already by Alexander, was then 
read through the prism of various other deities that seemed to possess 
analogous features. Though he was marginal among Alexander's gods, 
Serapis became important once Egypt came to be ruled by a Mace- 
donian king. Quite apart from the other aspects of his identity, the 
new god obviously needed an appropriate face and bearing, that is, a 
cult-image. Of course the dates offered by the literary tradition do not 
fit with such a scenario, but in my view that is not a serious objection, 
since everyone accepts that the information it provides is unreliable. 

The most plausible explanation for the disagreements in the sources 
is that all three of the early Ptolemies played a part in the institutionali- 
sation of the cult. Since the discovery by Alan Rowe in 1941—2 of the 
hieroglyphic and Greek foundation plaques of the Ptolemaic enclosure 
on Rhacotis, there can be no question that it was built in the third 
quarter of the third century BC by Ptolemy III Euergetes. 

This enclosure, consisting of a colonnaded court measuring some 
220 x 75m, housed a temple of Serapis with a cult-statue, flanked by 
a contemporary 'Stoa-like Building' whose purpose is unknown (Text 
fig. I). 87 It had two semi-monumental entrances from the north-south 
street R8 on the eastern long-side. These entrances are not placed sym- 
metrically in the outer wall but seem to be aligned in one case directly 
opposite an earlier 'T-shaped Building', in the other, to provide access 
to the so-called 'South Building'. The main enclosure is thus firmly 
dated. But the T-shaped Building and the South Building make clear 
that some effort at monumentalisation had been undertaken earlier. 88 
There is even some epigraphic evidence that points in the same direc- 
tion. An altar to Ptolemy II and Arsinoe (died 270 BC) was found on 
an earlier pebble-mosaic floor belonging to a building cleared to make 
way for the temple of Serapis, and seems to commemorate a chthonic 



87 On the history of the site, and especially the differences between the Ptolemaic 
building and the Roman replacement (built at some point between the fire of AD 181 
and the death of Caracalla in 2 1 7), see McKenzie, Gibson and Reyes 2004. The fact 
that the Stoa-like Building is set back from the building-line of the temple of Serapis, 
and has a different design, implies a) that it was built later than the T-shaped Building 
in front, and b) that it was not a temple of Isis. 

88 Both were demolished at the time of the Roman reconstruction, which is why 
virtually nothing can be said about them except that they existed. 



SYSTEMS OF BELIEF 



55 








SOUTH BUILDING 



Foundation 
plaques b^H 











Foundation 
^™ plaques 



50 M 



Fig. 1. Plan of the Ptolemaic wall-foundations of the Serapeum-complex 

at Alexandria. 



56 CHAPTER TWO 

cult in their honour, probably in association with Serapis. During his 
excavations, Rowe also found dedications to Isis and Serapis from the 
reign of Ptolemy I or early in that of his successor. 89 It is thus tempting 
to believe that these two buildings are all that was retained by Ptolemy 
III Euergetes of one or more earlier monumentalising phase(s). 90 Of 
particular interest is the fact that they were linked by an underground 
tunnel some 60m long and lm wide cut with great labour into the 
bare rock, and then lined with brick (Text fig. 2). This may imply, as 
Judith McKenzie has argued (2004, 83), that the two buildings were 
contemporary; but it certainly suggests that oracles and a certain degree 
of priestly prestidigitation were already associated with the cult before 
the building of the enclosure by Ptolemy III. 91 

Nowadays most scholars argue that the creation of Serapis involved 
drawing elements from a variety of existing deities whose cult could 
be exploited for dynastic ends. One suggestion is that, as Clement of 
Alexandria claimed, he was a fusion of Osiris and Apis, the bull that 
symbolised fecundity and was worshipped in the famous temple at 
Memphis. 92 Apis seems to have derived his link with the dead through 
his association with Osiris in the Saitic period. The god's name would 
reflect this fusion, Sarapis/ Serapis being a syncopation of Osiris + 
Apis. 93 The link with Osiris also suggested a further association with 
the Nile, its annual flood, and so the prosperity of Egypt. In other 
words, Serapis is likely to have been a hellenised version of an Egyptian 



89 Altar on mosaic floor: OGIS 725 - Bernard 2001, 34-6 no. 8 (faint, partly effaced); 
also Grimm 1998, 82 fig. 83; dedications to Isis and Serapis: SEG 24: 1 166; 1 167-8 = 
Bernand 2001, 19f. no. 2; 27f. no. 4, with Fraser 1972, 2: 385 n. 367. 

90 There are traces of a wall beneath the T-shaped Building (see fig. 2), and run- 
ning beyond the South Building, which implies still earlier construction, but there is 
no means of linking it directly to the Serapeum-phase. 

91 The well-known late-antique reports of mechanical apparatus and lighting effects 
in the temple refer to the Roman Serapeum that was constructed after the fire of AD 
181, and it is quite unknown whether they also existed earlier: Rufinus, Hist. eccl. 1 1 . 
22f. (tr. A.T Reyes in McKenzie et al. 2004, 106); Sozomen, Hut. eccl. 1 . 15; Socrates, 
Hist. eccl. 16. 2-5; Theodoret, Hist. eccl. 22; cf. E Thelamon, Serapis et le baiser du 
Soleil, Antichita Alto Adriatica 5 (1974) 243. 

92 Clement, Protrept. 4.48.6 Stahelin, p. 75.29-31 Marcovich. 

93 This was first argued by U. Wilcken, Urkunde der Ptolemderzeit, I: Papyri aus Unter- 
Agypten (Berlin and Leipzig, 1904—21) 77—89. In Varro's view the name was a syncopa- 
tion of soros, the Greek word for a sarcophagus, and Apis (ap. Augustine, De civ. Dei 
18.5). Oserapis is the form found in SB I 5103 = PGrMag XL (mid-IV a ), from the 
Serapeum in Memphis. 



SYSTEMS OF BELIEF 



57 



ALTAR (PTOLEMY II AND ARSINOE). 








50 M 



Fig. 2. The site of the Serapeum before the monumentalisation by Ptolemy 
III Euergetes, showing the tunnel linking the T-shaped Building and the 

South Building. 



58 CHAPTER TWO 

god of the dead, whose cult at Memphis was renowned in the period 
immediately preceding Alexander's conquest. 94 

This re-modelled Egyptian god was then read through an existing 
Greek deity Pluto, as was recognised already in the Hellenistic period 
by Heracleides of Pontus and Archemachus of Euboea, who are cited 
by Plutarch, De hide 27, 36 le. His outward appearance and many 
of his divine traits were borrowed from the god of the Underworld 
(Stambaugh 1972, 47ff). These were the basic ingredients, the ethnic 
background of each component being taken carefully into account. 
Features of other deities, such as Dionysus, as a god associated with the 
Underworld, came to be added in order to round out his identity and 
functions; in order to create a cult attractive to the Greeks, his worship 
also incorporated details borrowed from the Eleusinian mysteries. 

As the tutelary deity of Alexandria, Serapis took over some of the 
features of the Agathodaemon, the beneficent daemon in the form of 
a snake that, according to the legend, had died as a result of the work 
involved in founding the city (Ps.-Callisthenes 1.32.5-7). In expiation, 
Alexander put up a temple to the agathos daimon and the cult became a 
popular form of devotion, as is indicated by the quantity of evidence 
for it found in private houses. It was later taken over as a vehicle for 
other ideological interests, such as the imperial cult. 95 We cannot pin- 
point the precise moment at which these sacred snakes start appearing 
with the head of Serapis. However the parallel appearance of other 
snakes with the head of Isis and the minting of coins with the device 
beginning in the reign of Hadrian suggests that the syncretism between 
the Agathodaemon and Serapis began at much the same time. 96 This is 
all the more interesting because, exactly contemporary with Plutarch's 
De hide, it constitutes documentary evidence for syncretistic processes 
and their underlying intentionality: certain popular beliefs are instru- 



91 See esp. J.D. Ray, The House of Osarapis, in PJ. Ucko, R. Tringham and G.W. 
Dimbleby (eds.), Man, Settlement and Urbanism (London 1972) 699-702, on the cult at 
Memphis at the time of Nectanebo I and II, based on the excavations there of the 
1960s. The hieroglyphic version of the foundation-plaques of the Ptolemaic temple, 
found by Alan Rowe, gives Serapis' name as Osiris-Apis: Grimm 1998, 83, figs. 84a— b, 
d, f-g; Bernand 2001, 42f. no. 13, pi. 6.13; cf. R.L. Gordon, s.v. Sarapis, OCD' 1355f.; 
Dunand 1973, 1: 50. 

95 F. Dunand, Les representations de l'Agathodemon a propos de quelques bas- 
reliefs du Musee d'Alexandrie, BIFAO 167 (1967) 37f.; eadem 1981, 277-82; Sfameni 
Gasparro 1997, 81; 88-90. 

96 M. Pietrzykowski, Sarapis-Agathos Daimon, in de Boer and Edridge 1978, 3: 
959-66. 



SYSTEMS OF BELIEF 59 

mentalised in order to give expression to new forms of religiosity 
that directly benefit the politico-ideological interests in control of the 
society 97 

Ptolemy's interest in creating a cult acceptable both to the Egyptian 
and the Greek communities, so as to encourage social cohesion focused 
upon his own person, implied an effort of religious engineering involving 
refined managerial techniques (Scheer 2000, 260-6). Our sources imply 
that this technical knowledge was provided by ingenious theologians 
at the king's court. 98 On the Egyptian side, there was Manetho, who 
surely played an important role in translating the regenerative symbolism 
associated with Osiris into a form appropriate to the character of the 
new god, who was after all supposed to replace him as the consort of 
Isis. This involved a certain amount of re-thinking in order to fit him 
for his new, more suggestive, more polyvalent, role. This task required 
the services of a specialist in Greek religion, and more particularly 
someone familiar with the world of the Eleusinian mysteries who was 
capable of providing the raw materials needed for a successful religious 
fusion. The sources' claim that Timotheus, a member of the Eumolpid 
family at Eleusis, was chosen is thus wholly plausible. 99 His contribu- 
tion must have been to introduce into the cult of the Egyptian god 
key elements of the Eleusinian mysteries that might at the same time 
serve to hellenise Isis. 100 



97 J. Alvar, Isis y Osiris daimones (Plut., De hide 360D), in Alvar, Blanquez and 
Wagner 1992, 25 If. 

98 Plutarch, De hide 28, 362a; Tacitus, Hist. 4.83.2 is the only source to mention 
Timotheus. Welles 1962, 288 n. 83, rightly pointed out that no source actually says 
that Timotheus and Manetho created the cult of Serapis, as is usually said. However 
the fact that theologians of the two sides worked together certainly supports the idea 
that they, or other men of the same calibre, took an active role in the formation of 
the new cult. 

99 So rightly Burkert 1987, 37; 73; Scheer 2000, 263. Welles' claim to the contrary, 
that it may be possible to create a cult-statue or a cult, but not a god (1962, 284), is 
in my view baseless. A mere glance at the religious movements of the Principate is 
sufficient to show how unreasonable the claim is. Of course the process of construc- 
tion is complex. But there is always something that provides a distinctive character, so 
that, at a specific historical moment when social and ideological circumstances have 
prepared the ground, the new formation becomes acceptable. 

100 Cf. C. Schneider, Kulturgeschichte des Hellenismus (Munich, 1967-69) 2, 885f.; 
Dunand 1973, 1: 66fE; eadem, Le syncretisme isiaque a la fin de l'epoque hellenistique, 
in AA.VV, Le syncretisme dans les religions grec que et romaine. Colloque de Strasbourg 9-11 juin 
1971. Centre de recherches d'histoire de religions de l'Universite des Sciences Humaines 
de Strasbourg et l'Universite de Besancon (Paris 1973) 79-93. 



60 CHAPTER TWO 

I have already pointed out that, since he was Osiris' replacement, 
Serapis also possessed the traits of a deity of the Underworld. 101 This 
link was mediated through the connection established between him and 
the divinised bull Apis, which was the embodiment of the dead Osiris 
at Memphis. 102 All this is clearly due to the inspiration of the Egyptian 
priest Manetho. The god thus united the two cultures in himself. He 
was a god of the Underworld but also, and to no lesser degree, a god 
of (oracular) healing. These are the two aspects of his saving power. 103 
He regulates things in the world by virtue of his civilising suzerainty, 
but he also rules the cosmos, just like Zeus, and Egyptian Amun/ 
Amnion. 104 We encounter this role of Serapis repeatedly in inscriptions. 
An anonymous text in Greek from Quintanilla de Somoza (prov. Leon) 
in Tarraconensis, for example, invokes Eiq Zzvq Z(coxr|p?) £epcx7uq 'Icxco, 
One, Zeus, Serapis (the saviour?), Iao. 105 Through Apis, son of Re, the 
Sun-god, who is represented as a bull with the solar disk between his 
horns, he came to be associated with the Sun (Stambaugh 1972, 79-82), 
thus reduplicating his power of ordering the cosmos. As the heir of 
Osiris he is a god of fertility, symbolising the agricultural cycle: hence, 



101 Cf. Tacitus' comparison with Hades-Pluto and Dis pater: Hist. 4.83f. 

102 See nn. 92 and 93 above. Osiris continued to exist as an independent deity, though 
he had to share his religious functions with Serapis: he remained a god of death, ritual 
and myth, while Serapis became a deity who conveys his orders through dreams and 
grants the boons desired by those who pray to him. F. Dunand has however pointed 
out that at the temple of Osiris and Isis of Kysis in the Great Oasis (excavated by 
M. Redde in 1989), although the main cult image was of Osiris, the votives, in the 
form of small, locally-manufactured plaques, seem all to be of Serapis and Apis (1999, 
106-12). 

103 On Serapis and healing, see Stambaugh 1972, 75—78, and Chap. 4.4.b.ii below. 
There is a good example of this process at Emporion (Ampurias): E. Sanmarti, 
P. Castaner and J. Tremoleda, Emporion: un ejemplo de monumentalizacion precoz 
en la Hispania republicana. Los santuarios helenisticos de su sector meridional, in 
W. Trillmich and P. Zanker (eds.), Stadtbild und Ideologie. Die Monumentalisierung hispanischer 
Stddte zwischen Republik und Kaiserzeit. Kolloquium in Madrid, Okt. 1987 (Munich 1990) 
117—43; I. Roda, La integration de una inscription bilingue ampuritana, Boletin del 
Museo Arqueologico National 8 (1990) 79f; E. Sanmarti-Grego, Identificacio iconografica 
i possible atribucio d'unes restes escultoriques trobades a la neapolis emporitana al 
simulacrum del Sarapis d'Emporion, Miscellanea Arqueologica. A Josep M. Recasens (Tar- 
ragona 1992) 145-54; J. Padro and E. Sanmarti, Serapis i Asclepi al mon hellenistic: 
el cas d'Empuries, Homenatge a M. Tarradell, Estudis Universitaris Catalans 29 (Barcelona 
1993) 611-28; J. Alvar, Los santuarios mistericos en la Hispania republicana, in III 
Congreso Hispano-Italiano. Italia e Hispania en la crisis de la Repiiblica, Toledo, sept. 1993 
(Madrid 1998) 413-423. 

101 Bricault 2005b; on Ammon, see J. Leclant and G. Clerc, s.v. Amnion, LIMC 1 
(1981) nos. 141f; also InscrDel. 2037 = RICIS 202/0338 (Serapeum C). 
105 Garcia y Bellido 1967, 130 = SEG 32: 1082 ter = RICIS 603/0901. 



SYSTEMS OF BELIEF 6 1 

like Isis-Thermouthis, he carries the cornucopiae. From Osiris he also 
derives his Pharaonic traits as protector of the kingdom, with all its 
cosmocratic implications. At the same time, Serapis became the consort 
of Isis: this change of divine partner allowed them to be represented 
in specifically Hellenistic iconographic form. 106 They often appear as 
such on Hellenistic coins, and shared temples not only at Alexandria 
but also in many of the harbour-cities of the Mediterranean. 107 

Clement of Alexandria cites the well-known Stoic Athenodoros of 
Kana, i.e. Athenodorus Calvus, the teacher of Octavian, for the claim 
that a certain Bryaxis, not however the sculptor of the Mausoleum 
of Halicarnassus, was commissioned to create the image of the new 
god. 108 One assumption commonly made since W. Ameling first collected 
the then known material in 1903 is that the statue in the Ptolemaic 
temple on Rhacotis was of Serapis sitting on a throne accompanied by 
Cerberus, modelled on an established type associated with both Zeus 
and Pluto but topped with a kalathos, the unmistakable symbol of the 
fecundity provided by the Nile-flood (PL 2). 109 Recent work however 
has suggested that there were so many different Serapis-types current 
in antiquity that it is impossible to determine with any exactitude how 
the cult-statue in fact appeared. 110 Nevertheless, despite the variety 



106 E.g. Clerc and Leclant 1994, 1: 679-81 nos. 121-48; cf. Iran tam Tinh 1990, 1: 
771—74 nos. 130—83. Ptolemy IV Philopator and Arsinoe III founded a large temple 
to themselves and Sarapis-Isis on the main street of ancient Alexandria (now Sharia 
el-Horreya), 5 blocks north of Rhacotis: B. Tkaczow, Topography of Ancient Alexandria 
(Warsaw 1993) 80, site 27, map B; Bernand 2001, 53-6 no. 18. 

107 For the coins, see L.Bricault, Sylloge Mimmorum Religionis Isiacae et Sarapiacae (SMRIS). 
Memoires de lAcademie des Inscriptions et Belles-Lettres (Paris 2008); shared temples: 
Steuernagel 2004. 

108 Protrept. 4. 48.4-6, p. 74f. Marcovich. Ameling 1903, 85f. got rid of the prob- 
lem of Clement's citation by complaining of the 'naivete' of Athenodorus. Since he 
argued in favour of the ascription to Bryaxis (ibid., 97f), the topic has been discussed 
to death; Hornbostel 1973, 51-58 with p. 390 famously concurred. But this identifica- 
tion is now, mainly on chronological grounds, generally considered impossible; nor is 
there good reason to accept any part of Athenodorus' fanciful account of Sesostris 
and the Serapis-statue, which is the sole evidence for Bryaxis' involvement. It is far 
from certain that this is a genuine citation and not a spoof: none of the usual refer- 
ence-books acknowledge this material as authentic, e.g. J. von Arnim, s.v. Athenodoros 
no. 19, RE 2 (1896) 2045. 

109 Ameling 1903; cf. Clerc and Leclant 1994, 666; 668-70 nos. 1-19. 

110 Cf. Clerc and Leclant 1994, 689f. show that there are in fact numerous devia- 
tions in detail from the supposedly canonical type. One of the commonest types, esp. 
in terracotta, in fact shows the god standing (Tran tam Tinh 1983). The main point 
is surely that there are no surviving Hellenistic representations of the type: all known 
examples, in all media, date from the High Principate, cf. Schmidt 2005. 



62 CHAPTER TWO 

of types, Serapis is always represented with a Zeus-head surmounted 
by the kalathos (PL 3); this clearly was the Hellenistic substitute for 
the Egyptian iconography of Osiris. For her part, Isis took over that 
of Tyche/Fortuna, thus helping to fix her claim to sovereignty in the 
minds of her worshippers (PI. 4). The pair is sometimes represented 
as a 'family', with Horus-the-child, i.e. Harpocrates, often shown as a 
baby being suckled at his mother's breast, a Hellenisation of one of 
the commonest Egyptian statuette types. 111 

Despite his absence from the mythological cycle, then, and despite 
recognition of his recent origin, the cult of Serapis succeeded thanks 
at least partly to institutional backing 112 On the one hand, the men 
who put him together as a divinity of abundance well knew how to 
appeal to the needs in view of which he was created. On the other, 
it was the prestige of the Ptolemies (often indeed their territorial con- 
quests), and of Isis, that enabled his cult to spread through the eastern 
Mediterranean. 113 Serapis' polyvalence helped meet the politico-religious 
uncertainties of the Hellenistic period: a god of everything, omnipresent, 
director of the cosmos, lord of production and reproduction, bulwark 
of monarchy, succour of the little man — and to all that he added per- 
sonal command of the world beyond the grave. This was the kind of 
symbolic capital that a god now needed if he were to triumph. 114 



111 Tran tarn Tinh 1973 presents the range of types of Isis suckling (sometimes she 
suckles the Apis-bull); cf. idem 1990, 1: 777-79. 

112 Even in Egypt, Serapis succeeded in replacing older Egyptian deities such as 
Osiris and Min/Pan: G. Holbl, Ersetzt Sarapis altagyptische Goiter in der romischen 
Provinz Aegyptus? in H. Heftner and K. Tomaschitz (eds.), Festschriftfur G. Dobesch zum 65. 
Geburtstag (Vienna 2004) 601-07. At the Memnonion at Abydos, however, his oracular 
functions were taken over during the later Principate by Bes, cf. Bernand 1969, 525f. 
no. 131 (proskynema of the priest Harpocras). 

113 I am sceptical of the notion that the monarchy did not exert any political pressure 
to favour the cult. The proof is supposed to be the lack of fit between the pattern of 
dedications to Serapis and the political map of Ptolemaic conquests. Needless to say, 
I do not believe in Fraser's thesis that the cult spread 'innocently': P.M. Fraser, Two 
Studies on the Cult of Sarapis in the Hellenistic World, Opuscula Atheniensia, 3 (1960) 
47. See the arguments of Stambaugh 1972, 93-98 and Brady 1987, 11-41; the latter 
goes carefully over the evidence in favour of the idea that the Ptolemies supported the 
spread of the Egyptian cults in the Aegean. 

111 Quite who the worshippers were is another question. Dunand 1973, 1: 59 
emphasises his lack of success even at Alexandria, since the authors at court took no 
interest in him; Stambaugh, on the other hand, has no doubts, remarking simply, 'we 
know that he (Serapis) answered those prayers with satisfying frequency' (1972, 98). 
However the important recently-published inscription from Rhamnous shows that 
by c. 220 BC the koivov xcbv Zapaittaaxwv there (composed of Athenian citizens in 
garrison) was sufficiently influential to induce the Athenian general Apollodoros son 



SYSTEMS OF BELIEF 63 

b. The Myth of Cybele and Attis 

toutcc 8e eyeveto |iev ox)8e7iote, ecm 
8e dei, Kai 6 |a,ev vovq d|ia 7idvta opd, 6 
8e Xoyoq to |iev np&xa td 8e 8e-utepa 
Aiyet. 115 

Sallustius, Llepi tcov Gecov kou xox> 
k6o"|101) §4 (p. 8.14-16 Nock), cf. Julian, 

Or. 5, 170a 

In the case of the myth of Attis and Cybele, it is far more difficult to 
set out the myth synoptically since the surviving accounts differ wildly 
from one author to another, and there is no way of summarising or 
synthesising them that does not traduce one or other later tradition 
or variant. 

There seem to have been several versions of the myth already in 
Phrygia that are hard to disentangle. 116 Pausanias himself acknowledges 
the difficulties he experienced in obtaining information at the sanctuary 
of Attis and the Mater Magna in Achaean Dyme (7.1 7.9), and contented 
himself with re-telling the account of the elegiac poet Hermesianax, a 
contemporary of Alexander the Great. 11 ' As we shall see, this version 
is of little use for our purposes. The situation is no better if we look 
at the other classical authors who claim to provide coherent, that is of 
course heavily- edited, versions of an exotic myth. We do not possess 
any original Phrygian documents that might give us an insight into the 
changes they have introduced and at the same time provide authentic 
information about the cultural context of the narrative. 118 



of Sosigenes to sell them land for the construction of a temple of Serapis and Isis, 
and sufficiently wealthy both to decree him a golden crown in return and to have the 
resolution inscribed on a stele: SEG 49 (1999) 161 = RICIS 101/0502. 

115 "All this did not happen at any one time but always is so: the mind sees the whole 
process at once, words tell of part first, part second." 

116 See now the excellent critical survey by Bremmer 2004. On the belief that the 
Phrygians were a wise and ancient nation, see J.B. Rives, Phrygian tales, GRBS 45 
(2005) 223-44. 

117 Cf. S. Heibges, s.v. Hermesianax 2, RE 8 (1913) 823-28. 

118 Hepding 1903, 5—97 is an exhaustive collection of texts; a selection, with Ital- 
ian translation, in Sanzi 2003, 219—312; cf. Y. Dacosta, Initiations et societes secretes dans 
I'Antiquite greco-romaine (Paris 1991) 195ff Borgeaud 1996/2004 offers a critical discussion 
of the various traditions, which perhaps errs in the direction of trying to tidy up the 
Roman versions too much. Lancellotti 2002 has been much criticised for her claim that 
the Pessinus myth presented a sterile anti-king who subverts the normative model of 
dynastic legitimacy; perhaps the most measured account is that of Roller 1999. 



64 CHAPTER TWO 

For our purposes, we need to leave aside the Lydian version told 
by Herodotus (1.34-43) and Pausanias (7.17.9), according to which 
Attis died as a result of the attack of a wild boar sent by Zeus: it is 
hopelessly contaminated by the story of Adonis. 119 These versions may 
be of value in providing evidence for the diffusion of the myth in other 
contexts, but they are irrelevant for the issue of how the pair Cybele 
and Attis was constructed, and for their cult in the Roman Empire. 
For that purpose, we need to concentrate on the Phrygian version. 
This was the source of the variants found in Classical literature, which 
link the death of Attis to his self-castration. 'The myth' is in fact an 
amalgam of stories about a variety of gods artificially connected with 
that of Attis; the collection of sources put together in chronological 
sequence by Hepding makes this fairly clear. Though a construct, the 
various stories do nevertheless acquire a definite coherence around the 
figure of Attis. We have in fact to do with the symbolic projection of 
an integrated cultural system; the variety of versions is reflected in the 
variety of stereotyped scenes presented by the iconography 120 

The version that circulated most widely was probably the one nar- 
rated by Pausanias in the mid-second century AD. 121 An idea of the 
popularity of this version can be gained from the fact that in the early 
fourth century the Christian polemicist Arnobius offers essentially 
the same one, albeit padded out with details taken from a work on 
Cybele by a certain Timotheus (probably not to be identified with the 
Eumolpid). 122 Earlier authors echo this account. Catullus, for example, 
devoted his entire poem 63 to Attis, in a politico-cultural context hostile 
to new cults being introduced from the East, notwithstanding the fact 
that the Mater Magna, and apparently Attis, had for almost two centu- 



119 Cf. Panyassis of Halicarnassus frg. 27 Bernabe (= 25 Kinkel) = Apollodorus, Bibl. 
3.14.4. Bernabe assigns it to the fragments from uncertain works, cf. VJ. Matthews 
(ed.), Panyassis of Halikarnassos. Text and Commentary (Leyden 1974) 120-25. 

120 The iconography of the various scenes of the myth are usefully analysed by M.J. 
Vermaseren, The Legend of Attis in Greek and Roman Art. EPROER 9 (Leyden 1966). 

121 Perieg. 7.17.10—12, which explicitly states that it is yvcopiutirtara, most widely 
known. 

122 Adv. nat. 5.5-7; cf. R. Laqueur, s.v. Timotheos 14, RE 6A (1937) 1338. He is 
taken to be the Eumolpid by Burkert 1987, 73. On Arnobius, see H. le Bonniec, 
Arnobe, Contre les gentils, 1 (Bude) (Paris 1982) 7-93; C. Champeaux, Arnobe, Contre les 
pa'iens, 3 (Bude) (Paris 2007) i-xxxii; M.B. Simmons, Arnobius of Sicca: Religious Conflict 
and Competition in the Age of Diocletian (Oxford 1995). T. Zielinski, Les origines de la 
religion hellenistique, RHR 88 (1923) 173ff once argued that the mysteries of Attis 
were influenced by Eleusis through Timotheus; see the comments of Sfameni Gasparro 
1985, 33; Mora 1994, 116. 



SYSTEMS OF BELIEF 65 

ries enjoyed worship on the Palatine. 123 This presence of the Phrygian 
gods in the centre of Rome induced Ovid to devote a long passage 
of the Fasti to the Mater Magna, in connection with the grand April 
festival of the Megalensia,} 2 * It is in this context that he offers a version 
of the same myth to account for the fact that Cybele's most devoted 
followers castrate themselves (223-44). This account obviously reflects 
the interest aroused among Roman intellectuals by the cult in the late 
Republic and under Augustus. The version of Diodorus Siculus, which 
is heavily contaminated by the myth of Apollo and Marsyas, was writ- 
ten at much the same time (3. 58f). From all these materials we can 
construct a fairly satisfactory idea of the narrative. Arnobius' detailed 
version is the liveliest, and I follow it to a large extent here. 

Fired with lust for Cybele, Zeus masturbated and the sperm fell onto 
the rock Agdos, which gave birth to a hermaphrodite named Agdistis. 125 
This creature evinced such a lamentable inability to control its sexual 
and other urges that the gods decided to supplement the deficiency 
by emasculating it. Dionyus/Liber was commissioned to carry out the 
divine plan. 126 He put wine into the spring where Agdistis used to drink, 



123 Nauta 2004, 610-18; cf. T. Callejas, El carmen 63 de Catulo: la cuestion del 
genero literario, in Actes del 9 Simposi de la seccio catalana de la SEEC 1988 (Barcelona 
1991) 159-66; Harder 2004; Harrison 2004. 

124 Fasti 4.179-90, 337-72, cf. Inscrlt XIII.2 pp. 435-38. 

12:1 M. Meslin, Agdistis ou l'androgynie malseante, in de Boer and Edridge 1978, 
765—776, argues rather unconvincingly that the relation beween Attis and Agdistis 
was homosexual (in Arnobius, Agdistis is still masculine). There are many parallels 
to the theme of fecundation of a rock by divine sperm. In the so-called Hurrite-Hit- 
tite cosmogony, Kumarbi, the father of the gods, bites off the testicles of Anu, which 
makes him pregnant with triplets, one of whom is Teshub. Teshub seizes power from 
Kumarbi, who, in order to recover it, decides to create a divinity to oppose him. So he 
masturbates onto a rock, from which is born an anthropomorphic stone named Ulli- 
kummi, cf. W. Burkert,Von Ullikummi zum Kaukasus: die Felsgeburt des Unholds. Zur 
Kontinuitat einer miindlichen Erzahlung, Wurzburger Jahrbuch filr die Altertumswissenschaft 
5 (1979) 253—61. The sole classical reference to the nexus divine masturbation-divinity 
born from a rock occurs in [Plutarch], De fiuviis 23 A, where Mithra is said to have 
produced an offspring by this means. This is clearly a transformation of the birth of 
Mithras from a rock (see below). W.D. O'Flaherty, Women, Androgynes and Other Mythical 
Beasts (Chicago and London 1980) explores with great verve the religious and social 
functions of hermaphroditism in India. Alhough she makes a point of comparison with 
other mythologies, one is surprised to find no mention of Attis nor of the dismember- 
ment of Osiris, which might be thought relevant to her argument. Agdistis is clearly 
an example of her category of 'chaotic androgynes' (p. 332). Classical mythology is 
silent on the sexuality of Cybele, though it of course mentions the castration of Attis; 
in Hinduism, by contrast, Sati (Dakshayani), an aspect of Devi, copulates like Isis with 
the erect, albeit severed, penis (linga) of her consort Siva. 

126 Arnobius, C. gent. 5.6.1-5 = Sanzi 2003, 279, Cybele no. 39.2. 



66 CHAPTER TWO 

to such effect that it had to sleep off its stupor. He then tied its testicles 
to its feet so that when the thing awoke the sudden wrench tore them 
off There was much blood, and out of the blood-stained ground there 
grew an almond-tree, one of whose nuts was eaten (or simply secreted) 
by Nana, the daughter of the river Sangarius, and apparently one of 
the hypostases of Cybele. By this rather unusual means, Nana became 
pregnant with Attis. Being unwanted, the child was exposed, but this 
resort by Sangarius failed of its object because a goat-herd named 
Phorbas took care of him. When the boy reached adolescence he too 
worked as a herdsman, of such extraordinary beauty that everyone fell 
in love with him, until finally Agdistis, being now solely female and 
as such identified with Cybele, succumbed to the charms of her own 
flesh and blood. 12 ' 

Ovid's account allows us to establish links with figures from other 
myths. In his version Attis is simply a young Phrygian herdsman who 
breaks his promise of remaining faithful to the Great Mother by agreeing 
to marry la, the daughter of king Midas of Pessinus. Agdistis/Cybele, 
whom we must take to be the abandoned lover, plans her revenge. The 
only ancient version explicitly to state that Agdistis/Cybele and Attis 
enjoyed a sexual relation is that of Diodorus (3.58.4), who also says she 
became pregnant. It must be admitted that Agdistis' character is hardly 
consistent with the idea of chaste maternal affection. At any rate, I am 
inclined to argue in favour of a dual interpretation of Agdistis/Cybele 
that more or less faithfully reproduces the complexity of the relation 
mother/lover. 128 The next climax is the vengeance wreaked by Cybele 



127 This is the version of Pausanias, who attributes to Agdistis the role in the tragedy 
assigned by other sources to Cybele, cf. Mora 1994, 119. 

128 In my view, the well-known symbolon cited by Clement of Alexandria refers to this 
hierogamy: ek %\)\ma.vox> eipayov gK K\))j,|3dXo\) ejiiov, £Kepvoip6pr|aa, xmb xov Jtacrrov 
■U7ie8t)v. 'I have eaten from the tambourine, I have drunk from the cymbal, I have 
carried the sacred vessels, I have pushed past the curtain of the marriage chamber': 
Clement, Protr. 2.15.3 Stahelin, p. 24.1 If. Marcovich = Sanzi 2003, 266: Cybele no. 29.1. 
Firmicus Maternus, De errore 18.1 (— Sanzi 2003, 304: Cybele 49.2) cites the same 
symbolon, albeit with the final phrase substituted by yeyova \ix>o%t\^ 'Axxeax;, 'I am an 
initiate of Attis'. This variant may not be an error or distortion, for there may have 
been two (or more) versions, such that hierogamy was the final goal of initiation. The 
argument of P. Boyance, Sur les mysteres phrygiens: 'J'ai mange dans le tympanon, j'ai 
bu dans le cymbale', REA 37 (1935) 161—64 — idem, Etudes sur la religion romaine (Rome 
1972) 201—04, is completely different. He offers an idealist interpretation according to 
which we have to do with a musical rite of purification, where the food is imaginary. 
Though he does not explicitly discuss the final phrase, it seems clear from his general 
position that he takes the hierogamy to have been a mere dream-fantasy I discuss this 
passage further below, Chap. 4.3.d, on initiation in the cult of Cybele and Attis. 



SYSTEMS OF BELIEF 67 

at the wedding of Attis and la. The singing of the marriage-hymn 
has hardly begun when Agdistis/Cybele, furious with sexual desire, 
bursts in upon the ceremony and turns it into a blood-bath. All the 
men present are filled with maddening sexual desire: Attis picks up 
a flint and emasculates himself beneath a pine-tree (which becomes 
his symbol as a god [PI. 20]), while his followers, filled with the same 
madness, cut off their testicles in public. In her misery, the bride cuts 
off her own two breasts. From the spot where the blood spills on the 
ground, violets grow. 

Things quieten down. Agdistis/Cybele begins to regret the horror 
she has caused and implores Zeus to allow her lover to be restored to 
life. Zeus grants that Attis shall ever after remain in his tomb free from 
corruption and decay, manifesting his immortality by the fact that his 
hair continues to grow and his little finger moves, perhaps as a sort of 
epiphany of his penis (Loisy 1930, 93 n. 2). 129 That the tomb of Attis 
remained a powerful image in the cult is suggested by a recently-dis- 
covered defixio from Alcacer do Sal in Lusitania which addresses the 
megaron, the subterranean chamber that received the corpse of Attis 
(and where he evidently remains), and is likewise to receive the body 
of the person who stole the writer's property 130 The death of Attis is 
likewise thematised by an analogous recent discovery, a defixio from the 
joint sanctuary of Mater Magna and Isis at Moguntiacum (Mainz) in 
Germania Superior, which appeals to bone sancte Atthis tyranne to destroy 
a man named Liberalis per tuum Castorem, Pollucem. l?A The reference to 



129 According to Pausanias, the tomb was still shown to tourists at Pessinus in II P 
(1-4.5). 

Li0 Domine megare invicte, tu quiAttidis corpus accepisti, accipias corpus eius qui. . . I have used 
the text of F. Marco Simon, Magia y cultos orientales: acerca de una defixio de Alcacer 
do Sal (Setubal) con mencion de Atis, MHNH 4 (2004) 79-94. Marco Simon himself 
argues that megare is a mis-writing for megale, and translates "Gran sefior invicto", but 
to my mind he does not succeed in showing why a supposed femine form of [liyaq 
should have somehow crept into the title. One might also understand megare as an 
abbreviated form of megale(nsis), 'having to do with Cybele', but what we need is a 
reference to the grave. Given that the appeal is to a site that contains the body of Attis, 
it is surely relevant that ueyapov can also mean 'tomb' (LSJ 9 sense III. 2), a sense clearly- 
related to its meaning in the cult of Demeter, where it denoted a pit into which piglets 
were thrown to decompose (Pausanias 1.40.6); cf. Egelhaaf-Gaiser 2000, 216. I take it 
that the neuter megarum (only attested in Latin otherwise in CIL XIV 18; 19 = RICIS 
503/1220—1 with Bricault's note = a meeting-room for Isiaci at Ostia) has acquired a 
false vocative form -e by contamination. Despite the early date, invictus must allude to 
Attis' astral/cosmic status (see p. 38 nn. 38 and 39 above). 

131 AE 2004: 1026, with the commentaries of Blansdorf 2004; 2008, no. 2. One 
of the same group of texts likewise refers to the megaron as a place of death: rogo te, 



68 CHAPTER TWO 

Castor and Pollux here suggests that it was his dual character as a deity 
associated both with the heavenly gods and with the world below, a 
divine sort-crosser in fact, that made it natural to appeal to Attis in 
such contexts. 132 But the primary association is with death. 

The classic interpretation of the myth is that of Frazer, according to 
whom Attis is to be classified with Osiris and Adonis. 133 All three are 
to be regarded as vegetation-gods, whose very similar myths symboli- 
cally represent the annually-repeated regeneration of nature. 134 One 
of Frazer's show-texts, at least as regards Attis, was Plutarch's claim, 
around AD 120, that the Phrygians believe that their god sleeps during 
the winter and wake up again in spring {De hide 69, 378e). 13S Another 
was a passage of Firmicus Maternus, in the mid-fourth century AD, 
according to which the Phrygians understand their festivals as a thanks- 
giving for the harvest and the following seed-time; Attis is the harvest, 



Mater Magna, megaro tuo (i.e. the victim, Tib. Claudius Adiutor) recipias, et Attis domine, 
te precor . . . Blansdorf 2008 no. 8 (all the defixiones from the sacrificial area date from 
c. AD 70-130). Julian, Or. 5. 168c refers to the place where Attis was laid after his 
death as an avcpov, for which megarum must be a synonym (or possibly a term for a 
site within the penetrale that represented it). 

132 Blansdorf 2004, 55f. sees them as tokens of Attis' equivalence with Zeus/Jupi- 
ter. However the Dioscuri are already associated with Agdistis/Meter in Anatolia: 
E. Schwertheim, Denkmiiler zur Meterverehrung in Bithynien und Mysien, in S. Sahin, 
E. Schwertheim and J. Wagner (eds.), Studien zur Religion und Kultur Kkinasiens. Festschrift far 
K. Dbrner. EPROER 66 (Leyden 1978) 829 no. 4. The well-known Hellenistic 'initiation'- 
relief from Lebadeia (CCCA 2: 1 3 If. no. 432, pi. CXXVII) which shows a number of 
divine figures, including the Dioscuri, in company with enthroned Cybele, suggests that 
they could be thought of as having a special connection with transitions or changes of 
status, cf. Sfameni Gasparro 1985, 22f; eadem 1973, 141fE However their immediate 
association in the defixio from Mainz is with the grave; the Dioscuri appear repeatedly in 
Roman funerary contexts, evidently on account of their transition between two worlds: 
"Le contexte funeraire est, de tous ceux oil les deux divinites trouvent place, celui qui 
ofire la gamme la plus riche de sujets et de schemas" (F. Gury, s.v. Castores, LIMC 3, 
1: 608—35 at 631; cf. 2: 50 Iff). Recent work on Roman mythological sarcophagi has 
reminded us of the arbitrariness of the evocations of classical myth in the Empire, cf. 
Zanker and Ewald 2004: merely because the Castores came back to life in most Greek 
versions is no reason to argue for the association in later contexts. 

133 p razer 1914, 1: 261—317; 2, passim; also H.S. Versnel in Bianchi and Vermaseren 
1982, 906f 

134 Attis as a tree-spirit: 1914, 1: 277-80; Osiris as corn-god: 1914, 2: 96-107; as 
tree-spirit: 107-12. A good account of Frazer's comparativist methods and pre-occupa- 
tions in Csapo 2005, 29-67. In the same sense, Pettazzoni 1924/1997, 84f cites [Hip- 
polytus], Ref. haer. 5.9.8: XbjOXiGi 5e ocuxov . . . Opuya; kcci z^-oepov aidxt)V..., without 
mentioning the context (cf. n. 31 above). 

133 1914, 2: 40—42. In fact Plutarch opposes winter and summer (not spring) and 
says the Phrygians tote |xev KaxetivaaiJ.O'ix; tote 8' avepyepaeii; Pcckxeuovtr; avncp 
xeXotiCTi, 'celebrate his falling asleep in winter and his re-awakenings in summer with 
ecstatic junketings'. 



SYSTEMS OF BELIEF 69 

in that he is born from it, and his emasculation is the same as what the 
harvester does with his sickle when the grain is ripe, his death being 
commemorated by the storage of the grain after it has been gathered 
in. 136 Frazer, very much under the influence of Wilhelm Mannhardt's 
(1831-80) neo-Romantic theories about the primitive Teutonic tree-cult, 
argued this thesis at vast and learned length. 137 Insofar as it applies 
to Adonis, however, it was subjected in the early 1970s to a wither- 
ing critique by Marcel Detienne, who complained that Tauteur du 
Rameau d'Or n'a voulu voir que l'affinite d'Adonis avec la vegetation: 
si ce personnage passe le tiers de l'annee dans les regions inferieures 
et le reste sur la terre, n'est-ce pas le preuve qu'il incarne l'esprit du 
ble?'. 138 It now seems clear that at least Adonis' supposed connection 
with wheat was simply invented by Frazer, who abused the compara- 
tive method in order to get all three gods to resemble one another as 
much as possible, thus creating a series of equations similar to those of 
ancient writers, such as Plutarch, in an effort to impose order on the 
divine world. Detienne's pin-pointing of myrrh as the myth's central 
symbol suggests that the narrative circulates rather round two nodal 
oppositions, eating/ condiments and marriage/perfume. The world of 
scents naturally produced by certain herbs provides a kind of index of 



136 Amare terrain volunt (Phryges) fruges, Attin vero hoc ipsum volunt esse quod exfrugibus nascitur, 
poenam autem quam sustinuit volunt esse quod fake messor maturis frugibus facit. Mortem ipsius 
dicunt quod semina collecta conduntur, vitam rursus quod iacta semina annuls vicibus reconduntur: 
De errore 3.2. = Sanzi 2003, 303: Cybele 49.1, with the comments of Turcan 1982a, 
194 ad loc. 

137 W. Mannhardt, Wald- und Feldkulte, 1: Der Baumkultus der alten Germanen und ihrer 
JVachbarstamme (Berlin 1875); Frazer usually cites his posthumous Mythologische Forschungen 
(Strasburg 1884). Frazer's main discussion of the theme is to be found in The Golden 
Bough 3 5: Spirits of the Corn and of the Wild (London and Basingstoke 1912) 1: 131-70. 
Until Smith 1987, the thesis was at one time widely accepted as a simple fact, e.g. 
Pettazzoni 1924/1997, 84. As I have noted in the Introduction, however, there have 
been recent attempts to substantiate the structural pattern, albeit within a non-Frazerian 
framework: Mettinger 2001; Xella 2001. 

138 M. Detienne, Les Jardins dAdonis: La mythologie des aromates en Grece (Paris 1972) 12. 
Detienne's argument, and that of J.-P Vernant's important introduction (pp. i— xlvii), 
are brilliantly expounded by Csapo 2005, 262-76. We should however note that not 
everyone has been equally enthusiastic about this structuralist analysis: G. Piccaluga, 
for example, was extremely critical at the time: Adonis e i profumi di un certo strut- 
turalismo, Maia 26 (n.s. 1) (1974) 33-51, cf. Gordon 1979b, 301-04; more recently 
J. Reed, The Sexuality of Adonis, Classical Antiquity 14 (1995) 321-44 at 322fE As Csapo 
points out: "Structuralism's model of ideology is a totalizing system from which there 
is no escape ... even prostitutes joyfully celebrate their uselessness and social inferior- 
ity" (2005, 276). 



70 CHAPTER TWO 

the social order, thus marginalising (but in my view not totally discredit- 
ing) Frazer's vegetative cycle. 

In considering the relation between Attis and that cycle, we must also 
take into account his explicit connection with the pine-tree. According 
to the myth, it was beneath a pine-tree that Attis emasculated himself 
and died (see PI. 20), and this event was ritually reproduced at the 
festival of the Mater Magna on 22nd March, the day arbor intrat, when 
the dendrophori, whose duty it was, carried a pine-trunk into the temple, 
which then withered over the following weeks. 139 Moreover, as we saw 
earlier, pseudo-Hippolytus says that the Phrygians called Attis 'the unripe 
ear of grain'. 140 The series of rhetorical questions posed by Firmicus 
Maternus is relevant here, since he asks what seeds and grain have 
to do with funerals, or death, despair and punishment with love, thus 
inadvertently revealing the way his opponents interpreted the myth. 141 
Severe modern critics prefer to marginalise such references. 142 

Although of course I accept that a myth may be read at several 
different levels, and may also contain contradictions that reproduce 



139 Cf. Cumont 1929, 52-55; Vermaseren 1977, 113 and 120; 1981c. Vermaseren 
suggests that reliefs showing the death of Attis beneath the pine were commonly dedi- 
cated in Metroac temples; their iconography often disguises or softens the harshness 
of the event (1981c, 425f). The withering of the pine in the temple is used as the 
indicative image in one of the new 'judicial curses' from the joint temple of Isis and 
the Mater Magna at Mainz, c. 70-130: ita uti arbor siccabit se in sancto, sic et illi siccetfama, 
fides, fortuna, faculitas, "just as the tree will wither in the sanctuary, so may reputation, 
good name, fortune, ability to act wither in his case": J. Blansdorf, Magical Methods 
in the Curse-Tablets from the Sanctuary of Isis and Mater Magna in Mainz, in 
F. Marco Simon (ed.), Magical Practice in the Latin West: Papers from the International Conference 
held at the University of J^aragoza, 30th Sept.-lst Oct. 2005 (Leyden 2008) no. 18. I discuss 
the rituals associated with the death of Attis further in Chap. 4.3.e below. 

140 [Hippolytus], Ref 5.9.8; see n. 31 above. 

141 Vellem nunc mihi inquirenti respondeant, cur banc simplicitatem seminum acfrugum cumfunere, 
cum morte, cum fastu, cum poena, cum amore iunxerunt? Itane non erat aliud quod diceretur? (De 
errore 3.3, omitted by Sanzi). I consider the entire passage 3.1-5 extremely instructive 
(see also Turcan 1982a, 188-98). 

142 M. Meslin, for example, denies that they were originally agrarian cults (1985, 185). 
According to him, the initiate who dies and is reborn is not reproducing an ancient 
ritual based on the vegetative cycle but, in a far more 'religious' spirit, expressing a 
profound idea regarding death and resurrection. It is more serious for my position that 
even Tryggve Mettinger omits Attis from his category of ancient dying-rising gods, on 
the grounds that the "resurrection is no original part of the celebrations" (2001, 27; 
cf. 157f.) In these matters speculation is unavoidable, but I believe that there was both 
an original mythification of the vegetative cycle and a later re-elaboration of these 
ideas in the context of initiation into the mysteries, based on models inherited from a 
period when the old ideas, as their original meaning faded, were modified to fit them 
into a new semantic setting. 



SYSTEMS OF BELIEF 7 1 

those in the real world, I would like to propose what we might call an 
'integrative model' that takes the ancient interpretations to be mutually 
compatible — which of course does not mean that any one or other of 
them must be correct. Accepting the myth of Attis, say, as an account 
of a natural rhythm allows us to acknowledge an entire range of oppo- 
sitions whose function is to turn chaos into order. The myth traverses 
a range of ambiguities thereby defining the limits that determine this 
particular cultural production. Foremost among these ambiguities are 
death and the experience of life after death. The development of a 
discourse about the gods' ability to satisfy such human longings provides 
a means of escaping the shadow of death, or at any rate modifying 
its horror by means of new ambiguities. At the same time, the narra- 
tive deploys a number of categories that together set the terms of the 
distinction between divine and human, in such a way however as to 
assuage other anxieties concerning the inexplicable. Within the general 
framework of a hierarchy of living beings, the myth sets out what is 
acceptable within the given culture, assigning alternatives — in this case 
bestiality, aberration, anomalous reproduction — to the category of the 
Other, with all that such distinctions imply about the individual's moral 
duties towards others. In that respect, I follow the familiar modified 
structuralism of the School of Paris, viewing myth as a site where the 
oppositions nature/culture, normality/ anomaly, and the risks inherent 
in deviation from the rules, are given free rein. 143 At the same time, 
myth contains further information. 

For me, the legitimacy of Frazer's procedure lies in the fact that 
he tried to lay out the co-ordinates of what I have termed a system 
of belief. This system comprised: the cosmic order, which is a func- 
tion of the nature of the gods; a sense of what it is to live and be a 
human-being in the world; and a vision of the world beyond death. 
If we analyse the myths, we find an aetiological account of the solar 
cycle, and, as a function of that, the vegetative cycle. But this level of 
analysis is by no means incompatible with the others that I examine 
in this book, those of the system of belief or the system of values, or 
that of ritual. The social order, as analysed by Detienne and Borgeaud, 
belongs to the imaginaire and really is, at least in part, shaped by the 



113 Cf. P. Borgeaud, L'ecriture a" Attis: le recit dans l'histoire, in C. Calame (ed.), 
Metamorphose du mythe en Grece ancienne (Geneva 1988) 87-103; J. Podemann Sorensen, 
The Myth of Attis: Structure and Mysteriosophy, in idem 1989, 23-29. 



72 CHAPTER TWO 

feelings of human-beings in the world and by ethical norms, that is, the 
system of values. It is in this area, where there is plenty of space for 
differing interpretations of the myth, that the full range of its cultural 
possibilities becomes clear. 

In the particular case of the Phrygian cults, then, the myth gives us 
an insight into cosmic processes linked to the dual nature of the gods. 
The emperor Julian in fact interpreted the myth as an allegory of the 
creation of the world, with Cybele animating the demiurgic Logos 
(Attis) that generates living beings. 144 It is also, and with less interpreta- 
tive strain, a faithful reflection of the desire for an explanation of the 
vegetative cycle, just as we saw in the case of Isis. This being a text that 
lays claim to being aboriginal, it can also be interpreted in many other 
ways, for example, as an ill-fated life-cycle myth, from adolescence to 
maturity 145 That would imply a didactic intention. Other readings too 
are possible, such as those I intend to offer later, related to the moral 
and the eschatological orders of things. 

Nevertheless there is some evidence that seems to legitimate an 
interpretation of this divine pair as creating a cosmic order capable of 
triumphing over the disorder caused by transgressions in this world. One 
is a recently-discovered defixio from GroB-Gerau, on the main military 
road from Mainz to Lopodunum, capital of the civitas Sueborum Mcrensium 
on the right bank of the Rhine. This text, dated late-first/early-second 
century AD, appeals to Attis, as deum maxsime Atthis tyranne, together 
with the entire company of the Twelve Gods (totumque duodecatheum), to 
vindicate the writer for the wrong done him by a woman, Priscilla, who 
has betrayed the sacra of a man named Paternus (perhaps the writer) 
and married someone else. 146 Until the discovery of this text no one 



111 At Or. 5, 161c he argues that Attis is the material form of the fertile, demiurgic 
intelligence that brings everything into being; at 1 66d he says explicitly that Attis is the 
demiurge par excellence; at 1 75a he calls him the unmediated demiurge of the material 
world, cf. V Ugenti, Giuliano imperatore, Alia Madre degli del (Lecce 1992), who analyses 
the myth in terms of neo-Platonic allegory. 

145 As is done by J. Strauss Clay 1995. Catullus' version of Attis may resemble the 
Black Hunter, but we would do well to remember that our information about him is 
far more extensive, so that we are in a position to see that the explanation in terms 
of a life-cycle ritual is too limited to be totally satisfying. The transition certainly fails, 
but it does so because of the transgression of the rule against incest, which lies at the 
heart of the myth. In my view, any analysis of Catullus 63 that does not make this 
point cannot fail to be a distortion. 

146 AE 2004: 1006a-b; M. Scholz and A. Kropp, Priscilla, die Verraterin: Ein 
Fluchtafel mit Rachegeber aus GroB-Gerau, in Broderson and Kropp 2004, 33—40; 
colour photos in M. Reuter and M. Scholz, Geritzt und Entzvjfert: Schriftzeugnis.se der romischen 



SYSTEMS OF BELIEF 73 

would have credited that Attis could have been addressed in such a 
manner at this date, or have been appealed to in order to reinforce a 
claim to justice, to right a wrong done to the speaker. 147 More doubt- 
ful — at any rate more subjective — is the evidence of the Parabiago dish, 
now generally dated to the fourth century AD (PL 5). 148 Attis is seated 
beside Cybele in her chariot drawn by four charging lions, surrounded 
by symbolism suggesting that a cosmic order has been definitively estab- 
lished. 149 If the majority view of the date is correct, as it surely is, this 
image could be fitted into the context of the allegorical interpretation 
offered by Julian, which tells us a great deal about the historical shifts 
of interpretation in the cult of the Mater Magna: 

Immediately after the castration, the trumpet sounds the recall for Attis 
and for all of us who once flew down from heaven and fell to earth. 
And after this signal, when King Attis stays his limitless course by his 
castration (oxe 6 PaoiXeix; "Axxiq YaTncn tf|V &7teipiav 8ia xr\q e.Kxo[ir\q), 
the god bids us also root out the unlimited in ourselves and imitate the 
gods our leaders and hasten back to the defined and uniform, and, if 
it be possible, to the One itself (eiti 8e to (bpia|ievov Kod evoet8e<; Koci, 
eutep oiov te ecmv, otuto to ev dvatpexsw). After this, the Hilaria must 
by all means follow. 

Julian, Oral 5, 169c, trans. W.C. Wright 



And: 



But when after staying his limitless progress, he has set in order the chaos 
of our world through his sympathy with the cycle of the equinox, where 
mighty Helios controls the most perfect symmetry of his motion within 
due limits, then the goddess gladly leads him upwards towards herself, 
or rather keeps him by her side (e7iavdyei 7tpo<; eoanfyv r) 9eo<; aa\xev(x>q, 
liaXkov 8e ex£i nap' eaa)xfi). 

ibid., 171c2-7 



Informationsgesellschaft (Stuttgart 2004) 70f. no. 108a/b/c. The grammar is unclear: Attis 
and the Twelve gods, in the accusative, seem to be invoked to guarantee that the curse 
will be effective, i.e. we should understand^ at the beginning. The actual destruction 
is to be accomplished by the local deae. 

147 The text proves conclusively that Lambrechts' belief that Attis only became a 
god in the course of III P cannot be sustained (there are several other recent texts of 
the same kind, which I allude to in Chap. 4.3.b). That does not however necessarily 
indicate that his main point, the late date of the introduction of the Hilaria, must also 
be wrong. 

148 See the discussion of the date, n. 343 below. 

149 Some good remarks on the cultural and literary context of the lanx by Turcan 
1992a, 74 and 1996b, 397f., where he defends the view that Attis represents the 
soul. 



74 CHAPTER TWO 

The text from GroB-Gerau perhaps legitimates the inference that, 
whatever the contribution of neo-Platonist ideas to Julian's conception 
of Attis, his contribution to the maintenance of moral order was a 
much older theme. 

c. Mithras AM 

Nam omnes, qui in haec studia incumbimus, ex 
iisdem veterum fontibus nostra haurimus, 
eademque omnes refundere cogimur. 

P. della Torre, De Mithra eiusque tabulis 

symbolicis, in Monumenta veteris Antii . . . (Rome 

1700) 157-252 at 157f. 

Wenig ist, was wir von Mithras mit Sicherheit 
wissen. 

G. Zoega, Ueber die den Dienst des 

Mithras betreffenden romischen Kunstdenkmaler 

(1798-9) p. 94 (ed. F. Welcker, 1817) 

[Zoega] erkannte im Mithras, welcher den Urtypus 
des Lebens, den Urstier opfert, den grossen Vermitt- 
ler zwischen dem guten und bosen Prinzipe, Ormuzd 
und Ariman. Das Gute siegt. Der Stier unterliegt in 
der Welt der Erscheinungen, aber sein Geist kehrt 
in Ormuzd's Lichtwelt zuriick, und seine Hulle 
wird zum befruchtenden Keim des Lebendigen und 
Lebenerhaltenden. 

Friederike Brun, Romisches Leben 
(Leipzig 1833) 113 

In trying to study the specific beliefs of the cult of Mithras we encounter 
not merely the interpretative problems we have already seen, but others 
too, arising from the fact that there is no discursive literary account 
of the kind we were able to use for the Egyptian and Phrygian cults. 
Indeed, there is no firm evidence that there even was a proper myth 



150 'Mithras' is the conventional form of the name in English and German, on the 
authority of scattered epigraphic evidence, e.g. V 827 (the nominative form is hardly 
ever found of course). It was probably the usual form in Latin. However it is certain 
that both the indeclinable Mi9pa and the Ionic/Atticising form MiGprn; were current 
both in Anatolia and among learned circles in the Roman Empire (the earliest impe- 
rial example of indecl. M{0pa is Justin, Dial. Tryph. 78.6; of MiGprn; Plutarch, Pomp. 
24.7). Lhe form in all the Romance languages, including Spanish, is Mit(h)ra. On the 
Indian and Iranian forms, cf G. Bonfante, Lhe Name of Mithras, in Duchesne-Guil- 
lemin 1978, 47-57. 



SYSTEMS OF BELIEF 75 

of Mithras. From the wretchedly inadequate literary sources, we know 
that several authors wrote more or less extensively about the cult. 151 
Porphyry for example, tells us that a certain Euboulus wrote a treatise 
on Mithras in many books (i.e. long chapters). It has been argued that 
this work was used by the early neo-Pythagorean theologians Numen- 
ius and Cronius in the mid-second century for their reinterpretations 
of ancient wisdom. 132 On the supposition that Euboulus' work was 
already known at this period, he may conceivably have been close to 
the process of creating the Roman mysteries, which, at any rate on 
Roger Beck's hypothesis, may have occurred within the loose context 
of the dynasty of Commagene and the eminent persons connected 
with it, such as the astronomer Tib. Claudius Balbillus.' 33 The dating of 
Euboulus is however highly uncertain. 154 On the other hand, Porphyry 
himself says that the best writer on the mysteries of Mithras was a 
man named Pallas. 155 However, although these two writers were used, 
or at least cited, by Porphyry, there is no evidence that they narrated 



151 The fullest presentation of the Classical, late-antique and oriental sources 
remains that of Cumont, TMMM 2: 1—73, supplemented by the texts collected in 
F. Cumont, Les mysteres de Mithra 3 (Brussels 1913) Appendice. However Sanzi 2003, 
411—41 nos. 1— 27 provides a useful basic collection, using recent editions and with an 
Italian translation, including the (rather hopeful) texts of the hymns or logia from the 
Santa Prisca mithraeum, published in 1965, and the so-called 'Mithraic catechism' 
(BBerol. 21196). 

in'2 Porphyry, De abstin. 4.16.2; De antra 6; Jerome, Adv. Jovin. 2.14; for the mid-second 
century dating of Euboulus (and Pallas), cf. Turcan 1975, 23-43; Beck 1984, 2055. 
On Numenius, see M. Frede, Numenius, ANRWIl.36.2, 1034-75. 

153 Cf. Beck 1998a = 2004a, 31-44. I myself think the scenario quite implausible, 
though this is not the place to pursue the point. 

154 Although it might be tempting, with Cumont, to identify him with the contem- 
porary of Porphyry who was head of the Academy at Athens around the year 260 
CE, since this would provide a literary work to explain the sudden interest of the high 
Roman aristocracy in the cult that begins in the late third century, Turcan is rightly 
sceptical (1975, 38f). There were several Platonist philosophers of this name, and no 
work on Mithras is recorded for the Athenian Euboulus. It is the general nature of his 
work and the way he associates Mithraism with the magi of Persia that may indicate 
a link with the emergence of the new cult; but Porphyry's interest in him may equally 
have arisen from Euboulus having recently 'platonised' the cult. It is difficult to come 
down firmly on either side. 

155 It is commonly assumed that Pallas lived at the time of Hadrian (see the discussion 
by Turcan 1975, 39f.). But this date is simply an inference from Porphyry, who says that 
according to Pallas Hadrian abolished human sacrifice (De abstin. 2, 56). On this reason- 
ing, we could be sure that Livy or Clement of Alexandria lived in the fourth century 
BO The trouble is that Porphyry does not place the two writers in any chronological 
relationship (I see no reason to follow Turcan in thinking that Pallas wrote a commentary 
on Euboulus), so that it is arbitrary to suppose that Pallas must have been earlier. 



76 CHAPTER TWO 

a complete or coherent myth of Mithras. Moreover it is striking that 
his cult is completely absent from Clement of Alexandria, Arnobius 
and pseudo-Hippolytus, which might well imply that there was no 
documented narrative for them to get their teeth into. What Christian 
polemicists, such as Justin and Tertullian, picked up (and surely distorted) 
was Mithraic ritual practice. In the present state of our knowledge, the 
uncertainty remains. Nevertheless, there can be no doubt that there 
was some sort of narrative account of the god's doings, whatever its 
degree of antiquity. It must have described his experiences in a man- 
ner analogous to what we know from the iconography of the Mithraic 
reliefs with side-panels or bye-scenes that summarised the beliefs of the 
god's adherents (PI. 6). 

Many scholars have offered similes to give an impression of the dif- 
ficulty of reconstructing a, or the, Mithraic narrative in the absence of 
a solid textual basis. My suggestion is that we should compare it to the 
catechism of Pedro de Gante (c. 1479-1572), which consists mainly of 
pictures (which he called jeroglificos, hieroglyphics) by means of which 
for fifty years he taught the indigenous Aztec children at his arts and 
crafts school S.Jose de Belen the rudiments of the Christian faith. 136 
The comparison gains additional value from the fact that the Mithraic 
images also had an educative or instructional component, which helps 
to explain their relative uniformity all over the Roman world. 

We are thus obliged to reconstruct the narrative mainly from the 
iconographic sources, supplemented by one or two isolated references in 
Classical sources. As a result, we can hardly hope to do more than gain 
a rough outline of the story, accepting that we have lost the subtleties, 
conscious and unconscious, that figure in the literary documentation, 
such as it is. This is one of the reasons why the cult is often considered 
somewhat different from the other oriental cults. 15 ' 



156 On 13th August 1523, three years after the conquest of Mexico, three Fran- 
ciscans of Flemish origin arrived in Veracruz. After they had learned Nahuatl in 
order to preach the gospel to the Aztecs, two of them perished on Cortes' expedition 
to Honduras, but Pedro de Gante (Pedro de Mura = Pieter van der Moere = Peter 
of Ghent) survived and taught in Mexico City all the rest of his life, cf. J. Martinez, 
Herndn Cortes (Mexico City 1990) 92. A facsimile edition of Peter's text-book has been 
published: Catecismo de la doctrina cristiana con jeroglificos, para la ensenanza de los indios de 
Mexico (ed. F. Navarro) (Madrid 1970). The most sophisticated form of the method 
was that used by Domingo de la Testera. 

157 The problem is well set out by Sfameni Gasparro 1979a and 1979b, though 
we would do well to heed the critical remarks of M.V Cerutti, Mithra 'dio mistico' 
o 'dio in vicenda'?, in Bianchi 1979, 389—95. There are at present very few scholars 



SYSTEMS OF BELIEF 7 7 

The central theme of Mithraism is the tauroctony, the killing of 
the bull. The young god, in heroic pose (for that reason he is very 
occasionally represented heroically nude), holds the animal down and 
drives his dagger into its neck, at the same time wrenching the head 
up by inserting his fingers into the nostrils (PI. 6). At the same time, 
he prevents the bull from getting up by pressing his left leg hard into 
its back. It is clear from several standard details that the bull's death 
is connected with the coming of life. The animal's tail often ends in 
one or more heads of wheat, which clearly connotes the idea that the 
sacrifice is itself fecundative. 158 Then again, a scorpion attacks the bull's 
testicles, perhaps so as to obtain some of its life-force. The snake, which 
is often shown licking at the wound, seems to be a chthonic symbol, a 
sort of receptionist at the desk of death. 159 

Apart from this stereotyped central scene, which offers numerous 
minor, indeed minute (and quite unimportant), variants with respect to 
the god's stance and the bull's behaviour, we find a variety of figures 
either incorporated into the scene or (mainly) above and around it. 
The most constant of these are 1) the two young torchbearers, Cautes 
and Cautopates, who were sometimes clearly regarded as gods in their 
own right, since they are the object of vows, and votive-offerings; and 
2) Sol (top left) and Luna (top right), who signify the ordered rhythm 
of the cosmos (PI. II). 160 Less frequent are the signs of the zodiac, 
either in a circle around the scene, or in an arch above (PI. 6); their 
connotation is the same, albeit this time focused upon the annual path 
of the sun through the ecliptic. Very occasionally more elaborate reliefs 
show the four main winds, signifying the cardinal directions, and so 
once again the cosmic order; and the seasons. In addition to these, the 
so-called panelled reliefs and frescoes, known mainly but by no means 
exclusively from Germania Superior, offer a variety of secondary scenes 
that seem to illustrate moments from the life of Mithras, scenes that 



who would subscribe to J. Toutain's enthusiastic declaration at the beginning of the 
last century that the epigraphic evidence allows us to "write the history of the cult 
of Mithras": La legende de Mithra etudiee surtout dans les bas-reliefs mithriaques, 
in idem, Etudes de mythologie et d'histoire des religions antiques (Paris 1909) 229. Note the 
positivist use of definite articles. 

158 In one very early case, three ears of wheat emerge from the wound itself in 
place of blood (V 593). 

159 However the other animals that usually appear on these scenes, the dog, the 
raven, sometimes the lion, cannot be understood in this way. 

"'" On this stereotype, see R.L. Gordon, A new Mithraic Relief from Rome, Journal 
of Mithraic Studies 1.2 (1976) 171 = Gordon 1996 no. Villi. 



78 CHAPTER TWO 

are occasionally found reproduced in mithraea as independent votive 
statues or reliefs (PL 7). 

Cumont was the first to offer a coherent narrative on the basis of 
the iconographic material, thus turning himself into the real mythog- 
rapher of Mithras, and building up a picture of an entire theological 
system, whose principles can be rediscovered by means of the scientific 
method. 161 In the 1970s Cumont's entire reconstruction came under 
sharp criticism, a matter I come to later. At the moment I need only 
say that the following account, which essentially resumes Cumont, is 
extremely speculative. 162 Almost every statement that is not a straight- 
forward description of an archaeological monument may well be 
mistaken. 

Out of Chaos there emerged a god of Unlimited Time, identified 
with Aion, Saeculum, Kronos or Saturn, and sometimes as Fate or 
Destiny. This deity is represented as a winged male figure with a lion's 
head, encircled by the toils of a snake (PI. 8). In his hands, he holds 
his attributes, a sceptre, a key or a thunderbolt. This primordial deity 
created heaven and earth, who then produced Ocean. A 'holy family' 
was thus produced, the supreme triad of the Mithraic pantheon. Heaven 
(Caelus) was equated with Zeus-Jupiter, who at some point received the 
thunderbolt from his father Kronos-Saturn, thanks to which he suc- 
ceeded in establishing himself as supreme god, and then gave life to 
the other gods who live on Olympus (PL 6, left pilaster). Over against 
them was a dark world directed by Ahriman-Pluto, who, being like- 
wise the son of Unlimited Time, was also brother to Caelus-Jupiter. 163 



161 See TMMM 1: 159ffi and 2932, largely followed by Vermaseren 1963. 

162 There is a certain parallel between Cumont's methods and those of contempo- 
rary classical archaeologists such as Adolf Furtwangler (1853—1907), who aimed to 
reconstruct an 'archetype' (Classical Greek sculpture) by comparing as many surviving 
(Roman) copies as possible (cf. P. Rouet, Approaches to the Study of Attic Vases: Beazley and 
Pottier [Oxford 2001] 38f). Cumont however came to believe he actually possessed the 
appropriate 'archetype', in the form of the Zoroastrian sacred books. A.H. Anquetil 
du Perron had of course published a translation of the Avesta in 1771, but it and many 
other texts were becoming available in a new English translation, still unreplaced, by 
E.W. West and his collaborators (including James Darmesteter) in the series Sacred Books 
of the East, nos. 6, 18, 24, 37 and 47, under the general editorship of F. Max Muller: 
Pahlavi Texts (5 vols., Oxford 1880-97). Though he was careful to use the more general 
term 'Mazdean', Cumont thus wrote Parthian, but more especially Sasanian and post- 
Sasanian Zoroastrian dualism straight back into the Roman evidence. Cumont also 
possessed a copy of Darmesteter's great French edition of the Avesta (1892-93), now 
kept as part of the library he left to the Belgian Academy at Rome. 

16i Cumont was heavily influenced here by J. Darmesteter's Ohrmazd et Ahriman: leurs 
origines et leur histoire (Paris 1877). 



SYSTEMS OF BELIEF 79 

His troop of demonic beings, sometimes represented as snake-legged 
giants, tried to wrest the throne from Caelus-Jupiter, but were defeated 
and hurled back into the abyss from whence they had come (PL 9). 164 
These demons have access to this world and are able to act negatively 
on human-beings and impel them to do wrong. 

The basic means of purification are fire and water, represented by the 
lion and the crater that appear in many reliefs from Germania Superior 
and in the Danube provinces (PL 16, right of main scene). 163 They were 
considered divine elements, and thus played a central role in ritual. The 
winds too, since they were considered to have an influence on nature, 
were divine. The order of the cosmos, as I have said, was symbolised 
by the Sun driving his four-horsed chariot across the sky by day, and 
by the zodiaca signs, which allude to his annual path. If Sol was one 
of the most important Mithraic deities, his counterpart was Luna, who 
drives her biga, harnessed to two oxen, across the night-sky Finally the 
planets, understood as a series of seven (i.e. including Sol and Luna), 
were believed to be the guardians (tutelae) of the initiatory grades. 

The god who established himself in the Roman period at the centre 
of this system of belief was originally an Indo-European divinity. In the 
Rg-Veda (where the form of the name is Mitra) it is his duty, alongside 
Varuna, to maintain the order of the cosmos and to watch over the 
moral behaviour of mankind. The earliest documentary evidence for 
him (also in company with Varuna) occurs in a cuneiform treaty between 
King Suppiluliuma of Hatti and Shattiwaza/Kurtiwaza who was king 
of what was at that time the puppet-state of the Mitanni. 166 In Iran, 
his proper sphere was the social order: contracts, treaties, marriage, 
friendship and so on. He was a judge, and the terrible mailed fist of 
justice; he was the light that precedes the sunrise, but also day-light tout 
court; fire became his emblem because it was carried before him; he was 
lord of animal sacrifice and of the rain that causes seeds to germinate 
and plants to grow. All this is known from Tast 10 of the Persian Old 
Avesta, the hymn to MiOra (the Avestan form of the name). Composed 



161 Cf. Turcan 1988, 247f. 

"" Cf. Turcan 2004. A third purificatory element, as we shall see, is honey, associ- 
ated with ra Xeovxiicd, initiation into the important grade of Leo, Lion (Porphyry, De 
antro 15). 

166 Cf. P. Thieme, The 'Aryan' Gods of the Mitanni Treaties, Journal of the American 
Oriental Society 80 (1960) 301-17. The treaty exists in two versions, dated by A. Gotze, 
The Struggle for the Domination of Syria, 1400-1300 B.C., in CALL 3 2.2 (1975) [1965] 
19, to the final phase of Suppiluliuma's reign i.e. 1347—6 BC. 



80 CHAPTER TWO 

in the fifth century BC, i.e. during the Achaemenid period, the hymn 
is full of bloodthirsty descriptions of the ghastly punishments dealt 
out to the wicked by Mi9ra. 1C/ The festival of this Iranian god was the 
Mithrakana, one of the great feasts both of the Achaemenid court and 
in the rest of the empire, celebrated at the autumnal equinox. After the 
military conquest of the Achaemenid Empire by Alexander, the cult of 
Mi0ra survived only at scattered sites in Anatolia and Armenia (and of 
course farther East). Two such areas from which we have evidence are 
the kingdoms of Pontus and Commagene, some of whose kings were 
named Mithradates vel sim., m while theophoric names in Mithr- are 
relatively common all over Anatolia. 169 It is nevertheless impossible on 
present evidence to trace much if any direct continuity between the 
Achaemenid, or more vaguely Iranian, cult and the Roman one. 1 ' But 



167 The standard edition is Gershevitch 1959, with an excellent Introduction; exten- 
sive coverage of Indian and Iranian Mitra/MiGra in Hinnells 1975b; in Iran alone: 
M. Boyce, A History of ^proastrianism, V: The Early Period. Handbuch der Orientalistik 
1.8.1.2 (Leyden 1996); a simplified account in Merkelbach 1984, 9-39. 

168 E.g. Diodorus Siculus 31.19.7 (Mithridates), 19.22 (Mithrobouzanes). The founder 
of the dynasty was Mithradates of Cius, cf. J.G.F. Hind, Mithridates, CAH 1 9 (1994) 
130—33. A remote relative of the Kings of Pontus and Cappadocia was the Mithradates 
of Pergamum who assisted Julius Caesar in Alexandria in March 47 BC: Josephus, AJ 
14.127-36; BJ 1.187-93. 

169 See e.g. L. Zgusta, Kleinasiatische Personennamen (Prague 1964); R. Schmitt, Irani- 
sche Personennamen auf griechischen Inschriften, in Actes du VII' congres international 
d'epigraphie (Bucharest-Paris 1979) 137-52; M. Alram, Nomina propria iranica in nummis. 
Materialgrundlage zu den iranischen Personennamen auf antiken Miinzen. Das iranische Personen- 
namenbuch 4 (Vienna 1986); R. Schmitt, Iraniernamen in den Schriften Xenophons. 
Iranica Graeca Vetustiora 2 (Vienna 2002). 

170 As is well-known, Cumont postulated a direct translation of the Iranian cult, albeit 
syncretised with astrology, into the Roman world. Apart from Vermaseren 1963, the 
most Cumontian account is the earliest version of Turcan 2000 (ed. 1 in the Que sais- 
je? series, 1981a). A few years earlier Turcan could even write: "Mais d'ou Numenius 
tenait-il que Mithra etait demiurge? Nulle part dans la litterature mazdeenne il ne joue 
ce role" (1975, 78). (Turcan 1992a, 193-241 reveals considerable modification of his 
earlier views.) There have however been many alternative suggestions, of which we 
may mention G. Widengren, The Mithraic Mysteries in the Graeco-Roman World with 
special regard to their Iranian Background, in AA.VV, La Persia e il mondo greco-romano. 
Accademia Naz. dei Lincei (Rome 1966) 433—55; idem, Babaklyah and the Mithraic 
Mysteries, in Bianchi 1979, 675—696 (much waving of the now meaningless Dumezil- 
ian wand); A.D.H. Bivar, The Personalities of Mithra: Archaeology and Literature (New York 
1998), J.R. Russell, On the Armeno-Iranian Roots of Mithraism, in Hinnells 1975b, 
173—82 and M. Boyce and J. Granet, A History of ^oroastrianism, 3: 2f>rastrianism under 
Macedonian and Roman Rule (Leyden 1991). On the sacred sites founded by Antiochus 
of Commagene, cf. J. Wagner, Dynastie und Herrscherkult in Kommagene: Forschun- 
gsgeschichte und Neuere Funde, MDAI(I) 33 (1983) 177-224; J. Duchesne-Guillemin, 
Iran und Griechenland in der Kommagene (Constance 1 984); H. Waldmann, Der kommagenische 
Mazdaismus, MDAI(I) Beiheft 37 (Tubingen 1991); D.H. Sanders (ed.), Nemrud-Dagi: The 



SYSTEMS OF BELIEF 8 1 

it is worth remarking that by the first century BC Mithras had become 
identified both in western Parthia and in Bactria with the Sun itself, 
who sees all (form: Mihr), which would parallel what seems to have 
happened independently in Hellenistic Anatolia. 171 

Both literary accounts and the iconography tell us that Mithras was 
miraculously born from a rock. 172 The young hero, as a boy or even 
a baby, emerges from the rock already dressed in his Phrygian cap, to 
connote his 'oriental' origin, and carrying in one hand a lighted torch 
and in the other a sword or dagger. The flame of the torch stamps him 
as a solar deity, and as giver of light; the sword is the instrument by 
means of which he bestows life through the death of the bull. 173 The 
birth-scene thus anticipates his two crucial roles, as giver of light and 
life. On one unusual birth-scene, from modern Vetren in Bulgaria, shows 
the rock-birth framed by the torchbearers and Sol and Luna, exactly as 
in a tauroctony which may be similar kind of proleptic device. 174 The 
myth can in this sense be said to be recursive. It is consistent with this 
that the rock is not only the world but also the universe, recalling the 
cave where the death of the bull takes place. 

The birth of Mithras from the rock is the commonest of all Mithraic 
scenes after the bull-killing The fact that his birth brings light seems 
to fit with the hypothesis of a Mithraic cosmogonic narrative. 1 ' 3 In 



Hierothesion of Antiochus I of Commagene. Results of the American Excavations directed by The- 
resa B. Goell, 2 vols. (Winona Lake, Ind. 1995-6); J. Wagner (ed.), Gottkonige am Euphrat: 
Neue Ausgrahungen und Forschungen in Kommagene (Mainz 2000). On the supposed Mithraic 
pirates, see now C. Rubino, Pompeyo Magno, los piratas cilicios y la introduction del 
Mitraismo en el Imperio romano segun Plutarco, Latomus 65 (2006) 915—27. 

171 Parthia: Strabo 15.3.13, 732C. A. Momigliano, Alien Wisdom (Cambridge 1974) 
139-41, suggested that the source was Apollodorus of Artemita (I a ), who was himself a 
native of Parthia; Bactria (the coinage of the Kujan kingdom): Gnoli 1979, 730-34. 

172 No pagan source mentions the rock-birth, but several Christian writers, for what 
they are worth, allude briefly to it, e.g. Justin, Dial. Tryph. 70.1; Commodian, Instr. 1.13 
(acrostic); Firmicus Maternus, De errore 20.1; Johannes Lydus, De mens. 3.26. Vermaseren 
1951 is still the best discussion; more briefly, Merkelbach 1984, 96—98. 

173 Mithras once holds a handful of wheat in his hand instead of a torch (Schwert- 
heim 1974, 17 no. 11a, mithraeum II at Cologne), just as the bull sometimes has one 
or more ears of wheat emerging from his tail before he is killed. These are of course 
variants of the 'synoptic' views common in didactic art, where the desire to instruct 
shoulders the rules of naturalism aside. 

174 Clauss 2000, 64 citing N. Gisdova, Ex-voto dedie a Mithra decouvert recemme- 
ment au village de Vetren, arr. Pazardjik, Archeologie (Sofia) 3 (1961) 50f. 

175 Cf Guarducci 1979, 158f. in the context of V 498. Flames occasionally emerge 
from the rock at the birth, e.g. in the cult-niche fresco of Dura-Europos, mithraeum 
III (V 42 no. 5) and on the panelled relief from Nersae/Nesce, V 650 L3. The precise 
sequence of events is wholly opaque, however. The later promotion of (Zeus Helios) 



82 CHAPTER TWO 

mithraeum III at Poetovio in Pannonia Superior was found a mid-third 
century altar with a representation of Mithras born from the rock, and 
being helped out, or supported, by two figures who are clearly Cautes 
and Cautopates (PI. 10). 176 Above them is the figure of an old man 
asleep, who must be Saturnus, being offered a crown by a Victory. This 
might represent Saturnus having a premonitory dream foretelling the 
birth of Mithras (Merkelbach 1984, 98; 118). But it might also be an 
attempt to indicate the passing of an era, in a hypothetical Mithraic 
chronography according to which the reign of Saturn/Infinite Time 
gives way, or shall give way, to the reign of Mithras (one of whose epi- 
thets at Housesteads is Saecularis), a reign that grants victory over evil, 
and cosmic peace. 177 Another aspect of such a chronography would be 
the roles of Sol and Luna in producing an endless sequence of days 
and nights; of the planets, which in Mithraic iconography certainly 
sometimes represent the week-days (e.g. V 693, Bologna); and of the 
zodiac, connoting the sequence of months that make up the solar year. 178 
There can also be no doubt, as we saw in relation to the Trier rock- 
birth, that the sequence of the seasons was of central interest in the 
cult. 179 As for the future, there have been several attempts to imagine 
a Mithraic eschatology 180 



Mithras as demiurge and kosmokrator seems to have overlaid an earlier version in which 
he was a mere hero and came into a world where, for example, Cautes and Cautopates 
already existed, as well as the sword with which he was to kill the bull. 

176 V 1593 = Selem 1980, 131f. no. 92 = Merkelbach 1984, 376 fig. 140; cf. 1430 
B3 = Merkelbach 367 fig. 131, from Virunum. 

"' V 863f. with CM. Daniels, Mithras Saecularis, the Housteads Mithraeum and a 
fragment from Carrawburgh, ArchAel. 40 (1962) 105-15. 

178 Among the verses at S. Prisca on the Aventine is the beginning of a description 
of the zodiacal sequence, Primus [et] hie Aries astric[t]ius ordine currit: Vermaseren and 
Van Essen 1965, 213 no. 13. 

179 On all these temporal sequences in the cult, cf. Beck 1988, 17f, 54f; Ulansey 
1989, 98-103. 

lso ■pjjg Mithraic conception of time and its implications for eschatology was dis- 
cussed a propos the discovery of the Phaethon scene of the Dieburg relief (V. 1247 — 
Schwertheim 1974, 160f. no. 123a Riickseite) by E Cumont, La fin du monde selon 
les mages occidentaux, RHR 103 (1931) 29—96. Cumont linked the new monument 
with the Zoroastrian doctrine of the Frasegird and the metals of Celsus' 'ladder' (ap. 
Origen, C. Celsum 6.22). Gordon objected that Iranian eschatological ideas were unlikely 
to have been present in the cult at all — at any rate the presence of a Phaethon relief 
in a mithraeum was far too flimsy a basis for such a claim, for it completely ignored 
the interpretations of the myth of Phaethon current in the Roman world in II 1 ' (1975, 
237—44). Turcan argued that Celsus' ladder evokes the Platonic doctrine of the Great 
(Equinoctial) Cycle, a period of some 25,700 years. I incline to support Beck's view 



SYSTEMS OF BELIEF 83 

In the case of the well-known representation of the rock-birth from 
the mithraeum at Housesteads on Hadrian's Wall, the god is shown as 
bursting out of an egg within a highly unusual, indeed unique, ovoid 
zodiac. 181 This is a clear allusion to the influence of Orphism on the cult 
in the third century AD, which is underscored by a roughly contempo- 
rary votive text in Greek of the same date from the Emporium area by 
the Tiber in Rome, where the god is addressed as Zeus Helios Mithras 
Phanes (V 475 = IGUR 108). 182 By this period, at any rate, Mithras had 
become himself a demiurge and kosmokrator, just as Euboulus claimed 
he was (Porphyry, De antro 6). The point is confirmed by another Greek 
inscription from Trastevere in Rome, for many years lost and often 
considered a forgery, that has now surfaced in South Africa, and is 
certainly genuine. Mithras is there given the learned compound epithet 
daxpoppovToSaVcov, the god who is lord of the stars and of thunder. 183 
The link between birth from a rock, control of the growing season and 
mastery of the kosmos is supplied by an important monument from 
the Altbachtal sacral area in Trier that shows a baby Mithras, with one 
hand supporting a zodiacal circle with the six spring and summer signs 
and holding a large globe (the universe) with the other. 184 The pediment 
above represents, among other things, the order of heaven. 

It has been argued that the myth contained a second or doublet ver- 
sion of the birth of Mithras (e.g. Merkelbach 1984, 99f). This view is 
mainly based on an image on the complex relief from Heddernheim 
(Frankfurt a.M.), where Mithras' head and shoulders emerge out of a 
tree. 185 Analogous, though slightly different, representations are known 
from Syria: at Dura-Europos Mithras is clearly a baby, and naked; and 
at Hawarte (late fourth century AD), he is actually stepping from out 



that there is no reason why the doctrine of the soul's celestial journey should not have 
been Mithraic as well as neo-Platonic (1988, 81£; 1994, 29; 2006, 44-50). 

181 Cf. G. Ristow, Zum Kosmokrator im Zodiakus: Ein Bildvergleich, in de Boer 
and Edridge 1978, 985-87. 

182 Phanes was the Orphic divinity of limitless light born from a cosmic egg, that 
is, the primal stone, the embryonic world subject to the influence of the constellations, 
cf E Cumont, Mithra et l'Orphisme, RHR 109 (1934) 63-72 (still well worth reading); 
Guarducci 1979, 160-62; Jackson 1994. 

183 IG XIV 998 = IGUR 1 25 (omitted by V), with Gordon 2006, 1 80-89. For another 
Greek text, V. 463.1 = IGUR 194a,b (mithraeum of the Antonine Baths), an acclama- 
tion to Zeus Helios Sarapis re-cut for Mithras, see p. 162 below. 

184 V. 985 = Schwertheim 1974, 229f no. 190b, with Beck 1988, 39-42. 

185 V. 1083 = Schwertheim 1974, 67f no. 59a Al. 



84 CHAPTER TWO 

of the branches. 186 It might appear as though a tree has simply been 
substituted for the rock. Nevertheless there is a clear difference between 
rock-birth and the tree-scenes: in the latter, Mithras has neither torch nor 
dagger; at Heddernheim he is dressed in his usual adult 'Persian' dress 
(which is why Cumont interpreted the scene as Mithras looking out for 
the bull). This implies that these are not birth-scenes — an iconography 
that in the Mithraic context is linked with light and the death of the 
bull — but representations of the special relationship between Mithras 
and growing things, especially plants of agricultural value. 18 ' Porphyry 
mentions that the grade Persian had a special responsibility as "guardian 
of plants", i.e. it represented or figured one of the god's essential roles 
{De antro 16). 188 In other words, these images on the reliefs have not 
so much a narrative function, as we have so far been assuming, but a 
didactic or mnemonic one. On the other hand, an unique detail on a 
recently-discovered relief now in the Israel Museum shows Mithras as a 
baby drinking juice from a bunch of grapes (PI. 1 1). 189 Here the theme 
of the motherless, rock-born child has been fused with the theme of 
wine from the banquet scene, in order to answer a pressing question: 
how did new-born Mithras survive independently of women? At the 
same time, the drinking of unfermented grape-juice anticipates the 
mystery of the relation between the blood of the bull and the wine of 
the banquet of Mithras and Sol — and of the Mithraists who partook 
of the shared Mithraic meals. 

The Heddernheim panel may also however help to solve the puzzle 
posed by a panel on the Dieburg relief (V 1247), where three heads 



186 Dura-Europus: V 45 = Rostovtzeff et al. 1939, 111 with pi. XVIII.2 (fresco 
above the arcosolium, phase III c. AD 240); Hawarte: Gawlikowski 1999; 2000a; 
2000b; 2001 [2004], 2007. 

"" Cf. already N. Turchi, Le religioni misteriosofiche del mondo antico (Rome 1923) 187; 
Vermaseren 1963, 62. Priimm 1943/1954 actually called Mithraism a 'Naturkult'. 
Cumont rightly compared the youthful Malakhbel emerging out of a sacred cypress 
carrying a sacrificial kid on his shoulders on the reverse of the Capitoline altar 
(= Helbig 1 2, 30 no. 1182 with a misleading interpretation heavily dependent on du 
Mesnil du Buisson = Schraudolph 1993, 236 Cat. no. L145 with pi. 41 top, central 
photo) but wrongly understood this as connoting 'the solar god born from a tree' 
rather than as a reference to Malakhbel's connection at Palmyra with the sacred wood 
(ap. Rostovtzeff 1939, 11 If). He later realised that a similar motif was also used for 
Syrian Dusares. 

188 Vermaseren 1963, 124; Merkelbach 1984, 29 n. 22. 

189 See De Jong 1997, 55 with fig. 2 on p. 56; Gordon 2001a, lOlf. 



SYSTEMS OF BELIEF 85 

in Phrygian caps are found together on a tree. 190 These must represent 
Mithras with Cautes and Cautopates alike as masters of nature, recall- 
ing the expression xpinXaaioc, 9eo<;, threefold god, used in relation to 
Mithras by pseudo-Dionysius, Epist. 7.2 (that is, sixth century AD). The 
identity of the torch-bearers is not easy to determine, for they seem 
to have different values depending on the narrative context and their 
iconographic position. 191 Sometimes, as we shall see, they are simply 
the servants of Mithras, such as when they carry the dead bull back 
from the cave to the site of the first sacrifice, boil up its flesh (PL 1 1 , 
bottom left) or act as waiters at the feast of Sol and Mithras (PI. 15). 192 
Yet sometimes, as when they assist the birth of Mithras, they seem to 
pre-exist him; at other times, as when they drink the water he causes to 
gush out of the rock, they are the necessary witnesses of his miracles. 
In these roles they seem to have no obvious allegorical function. If 
we look at the tauroctony relief, however, we usually find Cautes on 
the spectator's left and Cautopates on the right of the main scene (PL 
1 1). In these positions, they seem to represent the rising/morning and 
evening/ setting sun, with Mithras, by implication, as the sun at mid- 
day 193 When they are in these positions, Cautes is also beneath Sol (day) 
and Cautopates beneath Luna (night), suggesting a further significance 
or association — Cautes is in fact fairly often represented with a cock, 
Cautopates once or twice, most clearly on the larger relief from the 
mithraeum of Castra Peregrinorum on the Caelian in Rome, with an 
owl. 194 When their location is reversed, however, i.e. Cautopates on the 



190 Cf. also the analogous scene at Poetovio II, where a head emerges from each 
of three cypresses (V 1510 = Selem 1980, 118f no. 70c = Merkelbach 1984, 372 fig. 
136 [good photo]). 

191 Cf. Hinnells 1976. The etymology of the two brothers' names seems to be Iranian, 
though the experts cannot agree upon a solution, and a Greek etymology has been 
seriously offered. Kaut- may be related to the idea of 'light', and the ending pat- signify 
'opposite', so the original meaning (in Median) might have been something like 'Light' 
and 'Anti-light', cf. M. Schwartz, Cautes and Cautopates: the Mithraic Torchbearers, 
in Hinnells 1975b, 406-23. 

192 Carrying bull: V 42.12 (Dura); carrying cauldron: de Jong 1997, 56f; servants: 
V 641 (Fiano Romano); 798 (Troia, Setubal); 988 - Schwertheim 1974, 239f. no. 206 
(Trier), cf. 635b (Rome). 

193 At Stockstadt II (Germania Sup.), they are associated respectively with East 
and West (V. 1212-14 = Schwertheim 1974, 148 nos. 117b and c). Kellens 1979, 
716 observes that the 'arms' of Avestan Mi0ra twice connote sunrise and sunset, and 
raises the possibility that this was the torchbearers' 'original' nature, whatever quite 
that would mean. 

191 Cock: Merkelbach 1984, 12 If; owl: ibid. 1 10, though as usual many of his sup- 
posed cases are imaginary. 



86 CHAPTER TWO 

spectator's left, Cautes on the right, as in about a quarter of all cases, 
a phenomenon that mainly occurs in the Danube area, they seem to 
connote the two seasons, Spring and Autumn (Pis. 6, 16). 

On an important monument from mithraeum III at Heddernheim/ 
Frankfurt, Cautes is further associated with Caelus, Heaven, and Cau- 
topates with Oceanus. 195 This suggests that the torch-bearers might 
also be identified with the opposition between life and death. This is 
the direction that Roger Beck in particular has pushed the analysis. 
He has argued that yet another role of the twin torch-bearers is as the 
agents of Mithras who are linked with the gates through which the 
souls descend into the world of genesis and re-ascend into immortality: 
according to Porphyry, De antro 24 (= Sanzi 2003, 422: Mithras no. 8.7), 
Cautopates is linked to North, cold and the genesis of souls, Cautes to 
South, heat and the apogenesis (release) of souls. 196 In these roles, they 
seem to represent, with Mithras, a sort of unexpected trinity, which 
may be the deeper significance of the panels at Dieburg and Poetovio. 
It is moreover certain, as I have mentioned, that the torch-bearers were 
considered gods of the kind one could make vows to in the expectation 
of receiving aid, and there are many named statues of them as gods 
in their own right. 

The central Mithraic narrative however focused on Mithras and the 
bull. The bull itself seems to have come from the Moon (Luna, as a 
planet, is the diurnal and nocturnal 'house' of Taurus, just as Sol is 
of Leo). The complex reliefs show a tangled sequence of events that 
include Mithras wrestling with the bull in order to overpower it, Mithras 
being carried off by the bull, clinging to its horns, Mithras riding the 
bull (PI. 12), and Mithras dragging the bull by its hind-legs (Pis. 7; 11 
bottom right), in the manner often found on Roman sacrificial scenes 
of smaller animals — only Mithras can drag a full-grown bull in this 
Appropriately for a rite of passage, these may be read as 



195 y 1 127 — Schwertheim 1974, 81 no. 61c. This may however simply be because 
they are two sets of brothers. 

196 Beck 1988, 93-95; 1994; 2000, 158-60; 2006, 107-12. He combines this with 
the apparent paradox that souls came into genesis through Cancer (summer solstice 
but the northern tropic) and left it {apogenesis) through Capricorn (winter solstice but 
the southern tropic). 

197 Merkelbach claims that this scene represents an initiate carrying the dead bull 
(1984, 343; 355; 358), despite the fact that the position of the tail and the forelegs 
make it quite clear that the bull is still alive. This is just one of hundreds of reckless 
hypotheses that make it necessary to use Merkelbach's book with great caution, despite 



SYSTEMS OF BELIEF 87 

allegories of the tests that human-beings have to overcome. A statue 
of Mithras in this pose at Mithraeum I, Poetovio (Ptuj, Slovenia), is 
labelled Transitu (m) , i.e. (a statue of) the 'Transition', from which we 
get a good impression of the effort required by the god to achieve his 
goal (PL 7). 198 

A raven gives Mithras a message from Helios/ Sol that he should kill 
the roaming bull. 199 He faithfully fulfils this task, as the tauroctony shows 
(PI. 6). 200 The scene has been interpreted as the moment of creation 
of all beneficial things, which makes Mithras into a true creator god. 
Indeed, Porphyry describes him explicitly as xov tccxvtcdv noit\xox> kocI 
naxpbq M{0pov>, Mithras the creator and father of everything (De antro 6 
= Sanzi 2003, 419: Mithras no. 8.3). And a little later on, in an obscure 
passage, he describes Mithras as Sruiioopyo!;. . . Kai yeveoecoq 8ean6xt\c„ 
demiurge and lord of generation (De antro 24 = Sanzi p. 422 no. 8.7). 
The images that show Mithras carrying a globe in his hand confirm 
this claim. 201 Cumont rather unconvincingly thought that the spirit of 
evil vainly sent poisonous creatures (the scorpion, the snake) to thwart 
Mithras' exploit. 202 However that may be, there can be no doubt of 
the total success of his mission, or that the killing of the bull was for 



its excellent photos; see the reviews by R. Turcan, Gnomon 58 (1986) 394-99 and R.L. 
Beck, Phoenix 41 (1987) 296-316. 

198 V 1494 = Merkelbach 1984, 369 fig. 133. The inscription is CIL III 14354 28 = 
ILS 4246. 

199 The raven has the function of a messenger in many different cultures, cf. 
J. Alvar, Materiaux pour l'etude de la formule sine deus sive dea, Mumen 32 (1985) 236—73 
at 253f. 

200 Cumont believed that he did so against his own will: "Mithra remplit a contre- 
coeur cette mission cruelle", TMMM 1: 303. The claim does not seem very plausible, 
particularly since in many cases he does look towards the wound or the bull's muzzle. 

201 Cf. V 334 = Merkelbach 1984, 298 fig. 42 (Altieri relief); V 985 = Schwert- 
heim 1974, 229f no. 190b (Trier), with Turcan 1975, 78 n. 127; 85 n. 172. The same 
interpretation could be offered of the scene in which Mithras seems to substitute for 
Atlas in bearing the cosmos on his shoulders, which is only represented in V 1283 = 
Schwertheim 1974, 184f. no. 141a scene 1 (Neuenheim) [it is not certain that V 1292 
= Schwertheim 1974, 192f. no. 148, L2 (Osterburken, see PL 6 here) represents the 
same idea], where he not only supports the world but is a genuine kosmokrator, cf. 
Clauss 2000, 87 with fig. 53. We might also cite V. 390 R2 in this connection, where 
Mithras stands in a curious pose between two cypress trees, one hand raised towards 
an arc, one lowered, which he identifies as Mithras kosmokrator. Richard Gordon 
suggests to me it may represent Mithras tracing the ecliptic as he stands on the line 
of the equinoxes, the meeting-points of the ecliptic and the celestial equator (cf. V 
1510 scene 3 = Merkelbach 1984, 372 fig. 136: 'Mithras separating heaven and earth', 
Poetovio II). This would amount to the same general significance. 

202 The problems with this theory are outlined in Hinnells 1975a. 



88 CHAPTER TWO 

Mithraists the central fact in the history of mankind; and that is the 
reason why it enjoys a privileged place in all mithraea. 203 

According to Cumont, Mithras had even before this managed on 
several occasions to scupper Ahriman's efforts to destroy mankind. On 
one occasion Ahriman caused such a terrible drought that his rival was 
obliged to intervene by firing an arrow at a rock or cliff; pure water 
came gushing out, thus saving those under his protection and making 
him into a god of waters. This motif was especially attractive to his 
worshippers on the Rhine-Danube frontier. 204 Thanks to one of the 
texts from the Santa Prisca mithraeum at Rome, published in 1965, 
the matter is now however understood quite differently. On a mid-third 
century altar from mithraeum III at Poetovio in Pannonia Superior, 
Cautes and Cautopates are clearly visible with Mithras as he fires his 
arrow (PI. 13). 205 This must remind us of S. Prisca line no. 4;fons con- 
cluse petris qui [gjeminos aluisti ne[c]ta[re fr]at[re]s, "O rock-bound spring, 
which fed the twin brothers with nectar". 206 There is no hint here of a 
symbolic reading of an element necessary to save people suffering from 
drought. If we consider that the 'water-miracle' is Mithras' first act after 
his birth, it may be that it is to be connected with Mithraic cosmogony: 
after the ordering of the cosmos and the growth of plants, he creates 
water from a miraculous spring, in the company of the same beings who 
attended his birth. That would have the advantage of integrating nar- 
rative, iconography and symbolism. For our present purposes it hardly 
matters whether the claim that the gushing water had the same divine 
qualities as the nectar that grants immortality to those who drink it was 
original or was only developed later. 207 C. Aelius Anicetus and his son 



203 Cumont, TMMM I 188; 305. The recent discovery of the late fourth-century 
AD paintings of the mithraeum at Huarte/Hawarte near Apamea in Syria has now 
proved that the theme of evil was present in the cult (see p. 196 below). 

204 Cf. Vermaseren 1963, 72; Turcan 1992a, 217; 2000, 56. 

205 V 1584 = Selem 1980, 130 no. 91 = Merkelbach 1984, 374 fig. 138 (right-hand 
photo). 

206 Vermaseren and van Essen 1965, 193 = Sanzi 2003, 439: Mithras 26.1; Merkel- 
bach 1984, 113. The reading of this line, like almost all the others at S. Prisca, could 
be made to appear much more doubtful than I have suggested by means of the square 
brackets. This is mainly due to the way that Vermaseren presented them. Almost every 
letter should in fact be dotted to represent guesswork. The reader must be warned 
that they are almost always reprinted (e.g. by Sanzi) in a very misleadingly sanitised 
manner. 

207 The parallel with Exodus 17.3—6, where Moses causes water to gush from a rock, 
which was later reinterpreted as the source of eternal life, has often been noted, cf. 
Merkelbach 1984, 114; Clauss 2000, 72. 



SYSTEMS OF BELIEF 89 

dedicated an altar at Aquicum (Budapest) where the torch-bearers flank 
neither the tauroctony nor the rock-birth but a large krater, evidently 
a symbolic substitute for the value of those scenes (V 1765). Manfred 
Clauss suggested that it contains the water from the rock (2000, 74), but 
several monuments imply rather that it contains the blood of the bull, 
drained at the sacrifice and collected in this vessel. 208 It may even be 
that it was also held in a sense to contain the bull's semen, or at least 
its fertilising power. In any case, there can be no doubt that the krater 
and its contents are among the key symbolic operators in Mithraism, 
full of connotations and allusions, and that they deserve more attention 
than has been accorded them up to now 209 

The nectar that feeds the twins may even be identical to the 'shed 
blood', sanguine fuso , mentioned by line 14 at Santa Prisca. The miracle 
of the water would on this reading have the same symbolic value as the 
blood and the semen, as the liquid that gushes from the Jons perennis, 
which Clauss has perceptively identified as Mithras himself (2000, 72). 
But that is not all. Porphyry states that the Mithraic krater symbolises 
the spring, which must be the spring of the myth: napa tk> M(9pa 6 
Kpaxrip dvxi ir\q nt\jr\c, tetcxktcxi (De antro 17 = Sanzi 2003, 421: Mithras 
8.5). If so, we may take it that the krater in the tauroctony symbolises 
the 'water-miracle', which would thus assume its true cosmogonic 
role, not as an episode in a supposed 'biography' of Mithras but as an 
authentic commemoration of the creation of water, taken as a symbol 
for all life-giving fluids. 

This symbolism however also has a soteriological dimension. Here 
we are forced to rely on the rather convoluted argument of Porphyry's 
neo-Platonist exposition of the Cave of the Nymphs in Odyssey Bk. 
13.103—12. He first points out that souls coming into being are moist and 
that blood is welcome to them, just as we give plants water (De antro 10). 
Moreover, the nectar that comes from the spring, or its hypostasis, the 



208 There is admittedly no incontrovertible evidence that the krater contained the 
blood, but its position on a number of Rhine-Danube monuments, such as the Brigetio 
plaque (V 1727) suggests as much. The fact that the snake often seems to be drinking 
from it (e.g. V 1306 = Schwertheim 1974, 203 no. 161, Fellbach nr. Stuttgart) rather 
than directly from the wound (as on e.g. V 75 = Merkelbach 1984, 279 fig. 18, Sidon) 
surely strengthens the inference that the krater contains the same substance that the 
snake feeds on when it leaps up to the dying bull. 

209 Cf. Gordon 1998, 235, distinguishing between at least two different Mithraic 
vessels, as well as between their ritual and their symbolic connotations. There can 
however be no question that the primary content of the Mithraic krater is wine, 
whatever symbolic or mythical meanings that liquid was given. 



90 CHAPTER TWO 

krater, is the liquid of coming-into-being. 210 However, he then encoun- 
ters a difficulty, for Homer says that the kraters and amphoras in the 
cave were full not of water but of comb-honey (Odyssey 13.105f. ~ De 
antro 15). 211 To account for this, he claims that the bees have deposited 
their honey in the Nymphs' cave as a source of incorruptibility; it is in 
this connection that he mentions the role of honey in the purification 
of initiates into the Mithraic grade Leo, since the initiate has to reject 
water because it is hostile to fire. Mention of the Lions reminds him 
of the grade Perses, Persian, whose initiates also underwent a ritual 
involving honey, but here the symbolism was different, for it signified 
to (pv>A,cxktik6v, their role as protectors of fruits and harvests. Moving 
rapidly on in this ceaseless chain of associations and assimilations, Por- 
phyry then observes that if honey is connected with the preservation 
of living things it is also connected to their procreation and generation, 
for which water too is required. He can thus triumphantly explain why 
Homer's bees should build their combs in kraters and amphoras, which 
are 'really' symbols of springs and fountains (De antro 1 7) and connect 
these with coming-into-being and going-out-of-being 

It is quite uncertain how much of all this we may take to be part of 
the Mithraic system of beliefs. But perhaps we can say that, Mithras 
having established himself as demiurge at his birth, his world was empty 
of life. In order to create a natural cycle he had to do two things: 
introduce water (fluids) and establish plant-life. The first is depicted in 
the scene of the 'water-miracle', the second in the tauroctony (PI. 14). 
When that had been done, the process once begun acquired a secondary 
or additional resonance: the primary fluid, water, mingled symbolically 
with the bull's blood/ seminal fluid, the medium through which human 
souls came into being and could enjoy salvation, and with honey, sym- 
bolizing incorruptibility, symbol of eternity. All three (or four) fluids 
are thus present in the krater, which itself can thus symbolise both the 
birth of Mithras and the tauroctony. 

I would indeed be tempted to go further in this sub-Porphyrian vein 
and link all this with the motif of three-fold Mithras. I have already 



210 In §13 he equates kraters and amphoras, as Homer seems to, an identification he 
repeats in §17. That being the case, the fact that in §15 he speaks only of amphoras 
does not disturb me. 

211 On Porphyry's allegorical method and arguments, see F. Bufiiere, Les mythes 
d'Homere (Paris 1956); R. Lamberton, Homer the Theologian: Meoplatonist Allegory and the 
Growth of the Epic Tradition (Berkeley and Los Angeles 1989). 



SYSTEMS OF BELIEF 9 1 

mentioned that the primary means of purification were fire and water, 
which form a pair like Cautes and Cautopates. Through the detail about 
the Lions, we can establish a further connection between fire and honey. 
The krater, which contains all these, would thus, just like the Mithraic 
cave, contain the solar trinity of Mithras, Cautes and Cautopates (who 
certainly formed some kind of unity, as the Dieburg tree attests), as well 
as the three life-giving fluids, blood, water and honey. Connotations of 
this kind surely explain the apparently odd reference in an inscription 
from Pax Iulia (Beja, Portugal) by a sodalicium of Bracaraugustans to 
a st[u]dium cum cratera, a meeting-house with a krater. 212 My discussion 
of the wider importance of the Mithraic krater should settle doubts 
about its Mithraic provenance; I would indeed claim that we have 
here a documentary corroboration of the importance of that vessel 
in Mithraic ritual. 

Cumont grouped other events with the water-miracle as showing 
the hostility of Ahriman, the principle of negativity, and the saving 
power of Mithras. He thought, for example, that the god had saved 
mankind from massive floods and fires, and that the human race had 
thus been able to recover demographically and populate the earth 
under the protection of its tutelary god {TMMM I 306). This aspect 
however received minimal attention in Mithraism, since, as A.D. Nock 
observed, "Mithras had no erotic mythology" (1937, 112), and women 
were excluded from the cult. 

Having established a world safe for mortals, Mithras could end his 
mission on earth. To celebrate this, he arranged for a farewell feast 
with himself and Helios as guests of honour, attended by Cautes and 
Cautopates (PI. 15). When they had eaten, the two friends mounted 
Helios' quadriga, which carried them up to the abode of the gods, 
where Mithras took his place as the protector of his faithful followers 
(Pis. 6, right pilaster, 5 panels from bottom; 16, lower register). 



212 V. 80 Ibis = AE 1984: 465, where doubts are expressed about its Mithraic prov- 
enance, reinforced by J. d'Encarnagao, Trabalhos de Arqueologia do Sol 1 (1986) 104 no. 339. 
Some of his points have been discussed byJ.M. Garcia, Religioes Antigas de Portugal 1991, 
450f no. 463; cf. J.C. Edmondson, Mithras at Pax Iulia — a Re-examination, Conimbriga 
23 (1978) 69—86. I think there are one or two analogies with the inscription from 
S. Gemini in Umbria published by Ciotti 1978, where the Leones have a leonteum cum signo 
et cetero cultu exornatum. The first word, a 'lion-place', enables us to infer the character 
of the Bracaraugustan st[u]dium; the remainder would be the equivalent of our cum 
cratera. Moreover — and this is very important — the relief from Troia (Setubal: V 798) 
shows a large crater at the centre of the banquet of Mithras and Helios, which is the 
mythical exemplar or prototype of the regular ritual institution in the cult. 



92 CHAPTER TWO 

In broad outlines, this is the canonical reconstruction established 
by Cumont, with the necessary provisos (and some digressions). As I 
have already mentioned, however, much of it is now considered ques- 
tionable. The first overall criticism was by Richard Gordon, at a time 
when Mithraic studies had developed sufficiently to risk alternatives to 
the Cumontian edifice, then seventy years old. 213 The most enduring 
gain of this criticism was the decoupling of Roman Mithraism from 
the 'Hellenised magi' who, on the model of Isiac cult, were supposed 
to have spread a genuinely Iranian cult around the Mediterranean. 214 
Some scholars then suggested a still more radical interpretation, which 
viewed the cult as a Roman creation, with no more substantial link to 
the Persian background than the name of the god. 213 The second gain 
was to reconsider the influence of neo-Platonist interpretation on some 
of the key ancient literary texts that tell us something about the cult, 
which has been the particular merit of Robert Turcan. 216 Thirdly, it 
was now possible to examine the iconography without the distortions 
introduced by Cumont's Iranising spectacles, and set it into its natural 
Hellenistic-Roman context. 217 



213 Gordon 1975. The first Mithraic Congress was held in Manchester in 1971; it 
is important to remember that just a few years before L.A. Campbell had published 
a far more completely Iranising interpretation of Mithraic iconography than Cumont 
had ever attempted, based on the Zoroastrian opposition between menog (the spiritual 
realm) and getig (the finite world) (Campbell 1968). It is now wholly forgotten except 
by those who, knowing nothing of more recent research on the cult, come across it 
on a library shelf. 

211 On the knowledge of Iranian religion, and in particular of the magi, shown by 
Classical authors, see the solid work of A. de Jong, Traditions of the Magi: Zoroastrianism 
in Greek and Latin Literature. RGRW 133 (Leyden 1997). The uncoupling of the Roman 
cult from the Iranian magi must be right, but we lack an analysis of Porphyry's (i.e. 
Euboulus and Pallas') attribution of the first mithraeum to Zoroaster. There is not a 
single epigraphic or other reference to Zoroaster in the western evidence, yet these 
authors, and Firmicus Maternus (see below), found such traditions somewhere. 

215 Vermaseren came round to this view, with some reservations, in M.J. Vermaseren 
1981a, apparently under the influence of R. Merkelbach, whose version of it only 
appeared somewhat later (1984, 153—61). Its main proponent is now Manfred Clauss 
(2000, 7fi, 2 If). 

216 Turcan 1975. Note however the criticisms of P. Athanassiadi, A Contribution to 
Mithraic Theology: The Emperor Julian's Hymn to King Helios, JThS 28 (1977) 360—71, 
though she has since retracted her view that Julian was to any significant extent influ- 
enced by Mithraism. 

217 Cf. Hinnells 1975a; 1976. R. Vollkommer, Mithras Tauroctonus. Studien zu einer 
Typologie der Stieropferszene auf Mithrasbildwerken, MEFRA 103 (1991) 265-81 has 
proposed a typology that includes the major transformations over time, though his 
premises are quite unacceptable. He is also responsible for the unhelpful and confusing 
entry s.v. Mithras, in LIMC 6 (1992) 583-626. 



SYSTEMS OF BELIEF 93 

The fourth result was the partial abandonment of Cumont's nar- 
rative approach in favour of an astrological interpretation of the 
tauroctony. It came to be widely believed that the iconography had a 
solely allegorical value, and could be read as a star-map of some sec- 
tor of the heavens, thus providing supposedly privileged insight into 
the world of Mithraic beliefs. But before I discuss the implications of 
this revolutionary interpretation, I would just like to mention that yet 
another new line of possible interpretation has been opened up by 
recent work on two Kurdish sects, Ahl-e Haqq and Yezidi: a central 
feature of their cosmogonies, which seem to be pre-Zoroastrian, is the 
killing of a bull by Mithras. 218 If this line could be pursued, it might 
support the hypothesis of Elmar Schwertheim and Roger Beck that 
the cult originated in the area of Commagene/North Syria or at least 
from eastern Anatolia. 219 

The idea that the tauroctony might be a kind of star-map goes 
back ultimately to the late-antique scholiast on Statius, Lactantius 
Placidus. Proximately, however, it was developed first by the well-known 
French Enlightenment figure Charles-Francois Dupuis (1742—1809), 
an outstanding champion of revolutionary principles and anti-clerical 
thinking (he published under the name 'Dupuis, citoyen francois'). In 
1795 Dupuis published a vast work in seven volumes, plus a volume 
of plates, entitled Origine de tous les cults, ou religion universelle, where he 
propounded the theory that all myths were basically astronomical. 220 
As was usual in his day, Dupuis treated all solar divinities as essentially 
the same, but gave a privileged position among them to Mithras. He 
explained the tauroctony as representing Sol at the equinox in Taurus, 
an allegory of the triumph of spring over winter. 221 Without knowing 



218 P.G. Kreyenbroek, Mithra and Ahreman in Iranian cosmogonies, in Hinnells 
1994, 173—82. The Yezidis have recently been in the news, as the victims of bomb- 
attacks upon their settlements in N. Iraq. 

219 Schwertheim 1974, 18; Beck 1988 (mainly fuelled by his interest in 'star-talk'). 
Both, as Beck oberves, were of course in fact picking up one of Cumont's later sug- 
gestions. 

220 Note already his Memoire sur I'origine des constellations, et sur ['explication de la fable, par 
le moyen de Vastronomie (Paris 1781). He later published a condensed version, Abrege de 
I'origine de tous les cultes (Paris 1798; 1822 3 ), which nevertheless still occupied 597 pp. 

221 Only two modern scholars have discussed Dupuis in relation to the mysteries of 
Mithras: Smith 1990 and, at considerable length, Swerdlow 1991. In fact, however, 
he was only extending an astral interpretation of Mithras that was quite standard in 
Late-Renaissance accounts, all of which started with the muddled information of Lac- 
tantius Placidus, on Statius, Thebaid 1.717—20, cf. R.L. Gordon, Interpreting Mithras 
in the Late Renaissance, 1: the 'monument of Ottaviano Zeno' (V 335) in Antonio 



94 CHAPTER TWO 

Dupuis' work, Karl Bernhard Stark (1824—79), a prolific author, whose 
important Systematic und Geschichte der Archdobgie der Kunst, the first volume 
of a projected Handbuch der Archaologie der Kunst, appeared only 
posthumously in 1880, suggested a thorough-going astral interpretation 
of three fragmentary reliefs found at Dormagen on the Rhine. However, 
his views were sharply criticised by Cumont, flush with the authority 
given him by the new armoury of scientific history, and the entire dis- 
cussion was dropped. 222 This is hardly surprising given the dominance 
in the early twentieth century both of historicist explanations and, in 
this field, of Cumont's views. The idea was only revived as part of the 
anti-Cumontian wave beginning in the 1970s. 

The basic idea, as Dupuis already argued, is that the tauroctony is 
not the culminating point of a mythical narrative but the representa- 
tion of a moment in the star-strewn sky from which all the possible 
cosmogonic and soteriological claims of Mithraism could be deduced. 
The textual basis for such a symbolic account of the tauroctony could 
indeed be found in Porphyry: 

For Euboulus tells us that Zoroaster was the first to dedicate a natural cave 
in honour of Mithras, the creator and father of all (eiq ti|if|V xox> 7tdvt(ov 

7toir)TOi3 Koci naxpoc, MiSpoi)) This cave bore for him the image of the 

cosmos that Mithras had created (eiKovoc qepovxoq awS xox> <J7tr|Xaio'u 
xox> Koa|io\), ov 6 Mi6pa<; eSruiioupYriae), and the things that the cave 
contained, by their proportionate arrangement, provided him with symbols 
of the elements and climates of the cosmos. So after this Zoroaster the 
custom prevailed among others too of conducting the initiations in caverns 
and caves whether natural or artificial (\iexa 8e towov tov Zcopodatpriv 
KpoariaavTOi; Kai 7iapd xoiq aXXoiq, 8i' dvtpcov kou am^Xaiav eix' cuv 
aikocpucov e'ixe x£ip07tovr)T(ov xaq xekexaq &7to8i86vou). 

Porphyry, De antra 6 = Sanzi 2003, 419: Mithras 8.3 223 

Elsewhere he tells us that Mithras' proper position is on the equi- 
noxes (ta> |iev o\)V M(9poc oikeicxv KcxQeSpocv tt|v Kara tocq iarmepiac; 



Lafreri's Speculum Romanae magnificentiae (1564), available only in electronic form on the 
Web at EJUS 4 (2004). 

222 K.B. Stark, Die Mithrassteine von Dormagen, Jahrbucher des Vereins von Altertums- 
freunden im Rheinlande 46 (1869) 1-25; cf. his Zwei Mithraeen der Grqfiherzoglichen Alterthumers- 

ammlung (Heidelberg 1865), the first publication of the Osterburken relief (pi. 6 here); 
Cumont, TMMMl: 202; 2: 387. The remains in question, V 1011-16 = Schwertheim 
1974, lift no. 8, had been found already in 1821. 

223 Included by des Places as a doubtful fragment of Numenius (frg. 60). 



SYSTEMS OF BELIEF 95 

vnexatp-v), a claim that has caused a good deal of disagreement among 
the proponents of an astral reading 224 

At the Mithraic conference in Teheran in 1974, the Yale Iranist 
Stanley Insler, working quite independently of Roger Beck, read a 
paper that linked an astral reading of the tauroctony with an iteration 
of Cumont's claim that the mysteries had originated in Iran (Insler 
1978). He understood the death of the bull as marking the end of 
winter and the approach of summer, an opposition he argued is also to 
be found in the Avesta. This does not seem a very telling parallel, but 
he also argued that the configuration of the constellations represented 
in the icon suggested a date in April. On the assumption that the great 
Iranian festival of MiOra, the Mithragan, was celebrated on 16th of the 
7th month of the old Iranian calendar (whose details are however very 
obscure), he arrived at the conclusion that in the first century BC it must 
have been celebrated in mid-April. Originally, then, the Mithraic icon 
celebrated the date of the Mithragan, which was also the traditional 
date of the beginning of spring in Roman peasant calendars. 223 This 
reading produced a stir at the conference, since it seemed to offer a 
way of recuperating a version of Cumont's story. 

In fact Roger Beck had already given a paper at the APA meeting 
in December 1973 identifying, like Insler, all the elements of the tau- 
roctony with specific constellations on or near a particular stretch of 
the ecliptic: the bull is Taurus, the dog Canis major, the snake Hydria, 
the raven Corvus, the scorpion Scorpius, the torch-bearers Gemini, the 
ear at the end of the bull's tail Spica. The lion that in the Rhine- 
Danube area often appears below the bull is Leo, and the krater near 
him is Krater (Text fig. 3). The implication of this reading would be 
that Mithras is closely linked to the constellation Leo, which is both 
the diurnal and nocturnal House of the Sun. 226 



221 De antra 24 = Sanzi 2003, 422: Mithras 8.7; cf. R.L. Beck, The seat of Mithras 
at the equinoxes: Porphyry, De antro nymphamm 24, JMS 1 (1976) 95—98 (repr. Beck 
2004a, 129-32). 

225 Alessandro Bausani agreed with the astral interpretation in general, having indeed 
been the first scholar since Stark to raise the possibility in print: Interpretazione paleo- 
astronomica delle stele di Triora, Bollettino del Centra Camuno di Studi Preistorici 10 (1973) 
1—19. But he argued that the motif was an adaptation of the ancient Mesopotamian 
theme of the lion attacking the bull, which would imply a date in the autumn. He 
anyway considered Insler's dating to the spring untenable: Nota sulla preistoria astro- 
nomica del mito di Mitra, in Bianchi 1979, 503-15. 

226 Beck's position was developed gradually: see the various essays now collected in 
2004a. In 2006 however he presents a substantially new line on many of these topics, 
which, although important, I have no space to discuss here. 



96 



CHAPTER TWO 



LU v a 


(ft 
Q) 
■D 


w 

Pleiad 




> 






3 .Q 




< V 

H 2 



/ 



< 

LU 




• < 



£. 



") 






C 



c 



■5 © 

.tj o 

« Q 

.a < 

a . 

Sh y 

O ~— ' 
_ 



3 









br 



SYSTEMS OF BELIEF 97 

Beck's main interest was in determining the period of the year 
represented by the tauroctony On the assumption that it depicts the 
part of the ecliptic between Taurus and Scorpius, he reckoned it must 
represent the constellations on the ecliptic in the southern sky at a 
moment when Taurus is setting in the West and Scorpius is rising in 
the East at sunset. Such a situation occurs in mid-August, the time 
of the harvest — indeed the harvest must be the core reference of the 
allegory. Insler, on the other hand, took the constellations to be those 
visible at the heliacal rising of Taurus. This fundamental disagreement 
between the two earliest proponents of the astral hypothesis meant that 
the hope that it would yield irrefrangible, 'scientific' meanings — and 
that was surely the secret of its appeal — was very soon dashed. 227 

While Beck was still building up his interpretation, a number of 
other hypotheses about the astral significance of the Mithraic icon 
appeared. The discussion turns on rather fine astronomical details that 
are not always easy to follow. Virtually all the participants agreed that 
the bull must be the constellation Taurus, and that the various other 
figures represent zodiacal constellations. Yet no one could agree on 
the identity to be allocated on these premises to Mithras. 228 Michael 
Speidel located all the constellations on the celestial equator; on this 
hypothesis the only major constellation missing was allegedly Orion 
the Hunter, which must therefore be identical with Mithras. 229 The 
Finno-Swedish scholar Karl-Gustav Sandelin suggested Auriga. 230 
The debate became really heated when David Ulansey intervened 
with a book neatly calculated to appeal to a wide public, The Origins 
of the Mithraic Mysteries. 27 ' 1 For him, Mithras was to be equated with the 
constellation Perseus. The origin of this identification was to be found 
in the lucubrations of a group of Stoicising intellectuals in the city 
of Tarsus in Cilicia. In brief, Ulansey used the fact that occasionally 
Cautes is shown holding a bull's head, and Cautopates a scorpion (e.g. 
V 2120, 2122, Sarmizegetusa) to argue that they were associated with 



227 Another expression of the desire for a 'scientific' basis for interpretation is the 
appeal by D.R. Small to linguistics, and in particular grammar: The Raven: an icono- 
graphic adaptation of the planet Mercury, in Bianchi 1979, 531—49. 

228 See more fully Beck 2004c, 242-5. 

229 M.P. Speidel, Mithras-Orion: Greek Hero and Roman Army God. EPROER 8 1 (Leyden 
1980). 

230 Mithras - Auriga?, Arctos 22 (1988) 133-35. 

231 Ulansey 1989; cf. already his Mithras and Perseus, Helios 13 (1986) 33-62. 



98 CHAPTER TWO 

the constellations Taurus and Scorpius. 232 The equatorial constellation 
midway between these points is Perseus, who must therefore be Mithras. 
Then Taurus and Scorpius must represent Mithras' 'seat', as Porphyry 
tells us, that is, the equinoxes. But during the Empire, as everyone knows, 
the equinoctial constellations were Aries and Libra; the equinoxes had 
not been in Taurus and Scorpius for two thousand years. This slippage 
has been known since Copernicus as the 'precession of the equinoxes', 
and for Ulansey it provided the clue to the origins of the mysteries. 
For precession had allegedly been discovered by the Greek astronomer 
Hipparchus of Samos in the second century BC, which he noted as 
a very slow secondary movement of the fixed stars, in addition to the 
familiar westwards rotation, which gave rise to the gradual shift of the 
equinoctial constellations. 233 The hypothetical Tarsian intellectuals (for 
whom there is not a shred of direct evidence) later interpreted Mithras' 
role as cosmokrator as his lordship of precession. This widely-publicised 
and indeed well-written thesis, which appeared among other places in 
the Scientific American, has attracted a considerable popular following, 
but also been met with sharp criticism from the specialists. 234 As one 
reviewer put it: "In several ways . . . the author arouses admiration for 
his ingenuity but not the conviction that he is right". 233 



232 This association at least is a fact, however we explain it: see the material col- 
lected by Vermaseren in the context of the rediscovery in Brazil of the 'monument of 
Ottaviano Zeno' (V 335): Mithriaca IV: Le monument d'Ottaviano ^eno et le culte de Mithra 
sur le Celms. EPROER 16.4 (Leyden 1978) 36-41. 

235 Ulansey 1989, 76—81. The precession is in fact caused by the slight shifting of the 
earth's axis at an annual rate of 50.3 seconds of arc per year =1° every 71.6 years, so 
that the rotation axis of the earth moves round the ecliptic north pole roughly every 
25,700 years (the 'Great Platonic Year'). This shift is due to the differential gravitational 
pull of the Sun and the Moon and the fact that the earth is not a perfect sphere. As a 
result, the equinoctial points and the celestial pole move, which gives rise to the phe- 
nomenon of precession, most easily perceptible in the westward shift of the zodiacal 
sign in which the vernal equinox occurs (5,000 years ago it was in Taurus, now it is in 
Pisces). Its climatic effect is that summers and winters become progressively milder or 
more pronounced every 12,900 years (the Milankovitch cycle), because the positions 
of the equinoxes and solstices change slightly each year. Over the next 10,000 years 
or so, irrespective of global warming, winters in the northern hemisphere will thus 
gradually become longer and more severe. 

234 E.g. J. Duchesne-Guillemin, CRAI 1990, 281-85; idem, del. Bulletin de la Societe 
Astronomigue de Liege 52 (1990) 197-208; Swerdlow 1991; M. Schutz, Hipparch und die 
Entdeckung der Prazession, EJMS 1 (2000); idem, Hipparchs Deutung der Prazession: 
Bemerkungen zu David Ulansey, EJMS 4 (2004); reply by D. Ulansey, Once Again 
Hipparchus and the Discovery of the Precession, EJMS 3 (2003). 

235 J.G. Griffiths, in CR 41 (1991) 122—24. For an enthusiastic review, however, note 
I. Huld-Zetsche and M. Koppf, Germama 74 (1996) 291-98. 



SYSTEMS OF BELIEF 99 

Though it was for a time fashionable, the astral hypothesis has not 
succeeded in convincing many scholars. For one thing, it is very dif- 
ficult to date the time of year signified by a particular rising or setting 
phenomenon. Then again, does not the sheer number of incompatible 
hypotheses undermine the whole approach despite its superficial plausi- 
bility? According to Beck, the chances of a coincidence occurring like 
this (between the arrangement of the stars on this part of the ecliptic 
and the Mithraic icon) are extremely small; but in order to explain 
what the point is, he now asks us to follow him in an elaborate and 
impressively-constructed theory of 'star-talk' (Beck 2006). It is never- 
theless tempting to join Swerdlow in his radical dismissal of the entire 
line of argument (Swerdlow 1991). 236 Whatever the truth here may be, 
we can at least agree that Mithraism was influenced by astronomical 
knowledge to a far greater extent than any of the other mysteries: its 
adherents were evidently stimulated to an interest in how the universe 
is ordered. 23 ' The present tendency at any rate seems to be to envis- 
age a process, necessarily speculative in view of the absence of reliable 
data, involving an Iranian basis, its Hellenistic transformation, and a 
constant Roman adaptation and re-interpretation, the iconography 
serving as a recursive point, a constant stimulus to reinterpretation 
and new meanings, including the exploration of the implications of 
the belief that this was a Persian cult. 238 It is indeed the polyvalence of 
Mithraic symbolism that makes its interpretation so difficult, but also 
gives it a remarkable coherence, if we regard the tauroctony both as 
a pseudo-historical narrative of universal scope and as a sort of star- 
chart representing the recurrence of the equinoxes, 'the proper seat 



2!,) I do not feel competent to enter into the technical debate, but his criticism of 
Ulansey seems valid. However I would dispute Swerdlow's claim that Mithraism was 
a very limited sort of religion and represented no danger to Christianity, not so much 
because of the surprisingly superficial positivist arguments offered but because of the 
powerful impression the cult made in its own time. It is this that encourages me to reject 
his notion that it was just a sort of off-duty club, and view it as a complex phenomenon, 
both as regards its claims and its rituals, of religious, intellectual, historical and social 
importance, in which many people in the Empire found the spiritual nourishment they 
sought, even if the Church Fathers did not care for it. 

2!/ B. Jacobs, Die Herkunft und Entstehung der romischen Mithrasmysterien. Uberlegungen zur 
Rolle des Stiffen un zu den astmnomischen Hintergriinden der Kultlegende (Constance 1999) is 
a good summary of the issues. He uses the astronomical arguments to argue for an 
origin at Rome. 

238 Cf. R.L. Gordon, Persaei sub rupibus antri: Uberlegungen zur Entstehung der Mithras- 
mysterien, in: Ptuj im romischen Reich/Mithraskult und seine %eit: Akten des intern. Symposion Ptuj, 
11-15. Okt. 1999. Archaeologia Poetovionensis 2 (Ptuj 2001 [2002]) 289-301. 



1 00 CHAPTER TWO 

of Mithras' according to Porphyry [De antro 24), thus simultaneously 
conveying two aspects of the Mithraic understanding of Time. 

I think we can summarise the central meaning of Mithraic iconog- 
raphy as follows. The bull represents the idea of generative potential. 
It either represents, or is closely linked to, the Moon, understood as 
a source of life, which only the Sun can activate. The gods share an 
existence in heaven (PI. 1 7); one of them, Mithras, is destined to pro- 
duce the vital spark, since he is Sol Invictus. He steals the bull-Moon 
from its celestial home, represented by the temple (PI. 16, top) and, 
despite its resistance, kills it. The quarrel between Mithras and Helios 
perhaps derives from this act: Helios must have been the owner of the 
bull Mithras stole in order to bring about the creation of the world. 
They are reconciled however and shake hands (PI. 18); this may be 
the mythical reference of a ritual act between initiate and Mithraic 
Father in an otherwise obscure Greek symbolon recorded by Firmicus 
Maternus: Mvoxa (3ookA,071it|(;, owSe^ie rcaxpoq ayawu, "Initiate of 
a bull- theft, hand-clasper of an august Father" [De errore 5.2 — Sanzi 
2003, 432: Mithras 19.1). In this form, the phrase was an acclamation 
of a neophyte, syndexios being the technical Mithraic word for an initi- 
ate, but the act seems clearly to be modelled on the bye-scene where 
Mithras and Sol shake hands, an event that takes place after the killing 
of the bull and apparently before the banquet they share. 239 

The theft and killing was thus a deliberate, positive act on the part of 
Mithras performed in order to achieve the desired end: the bull's death 
is no mere sacrifice but the truly heroic culmination of a strange form 
of hunt. At the moment the death occurred, the world we know was 
created. This occurred mythically at a particular time of year, when 
the Sun's heat drives away the winter cold and brings about the regen- 
eration of nature. The creation of the world thus models the annual 
re-awakening of nature. Taking Mithras' 'proper seat' into account, 
we might argue that his birth {genesis = naturd) occurred at the autumn 
equinox, and the eternal liberation of the forces of nature (apogenesis) 
at the spring equinox. 240 



239 F; rm i cus Maternus introduces the symbolon by saying: Virum vero abactorem bovum 
colentes sacra eius ad ignis transferent potestatem, '[The Persians and Magi], who worship a 
man who steals away cattle, interpret his cult as (celebrating) the power of fire'. He 
claims that we know this because it is stated by propheta eius, evidently referring to the 
Mithraic claim that the mysteries had been founded by Zoroaster. 

210 This would be the logic of the dominant polarity as understood by Beck and others 
(Beck 1998b), though I rather doubt that Beck would follow me here. I accept that the 
suggestion in the text is pure speculation, for which there is no ancient evidence. 



SYSTEMS OF BELIEF 101 

We have no idea how the hunt-sacrifice was linked to the creation of 
mankind, but it is clear that the once-for-all act of creation made life 
possible and that this demiurgic act entailed or modelled the annual 
repetition of the process, so that it could subsequently be read at two 
levels, astrological and naturalistic. 241 The bull's death was not due to 
gratuitous violence but was necessary in order to initiate the cycle of 
life. It took meaning within a context of ideas associated with the god, 
as lord of the rhythms of life, already in the fifth-century BC Yast 10 
(the Hymn to MiOra). The cosmogony linked Helios-Sol to the cre- 
ation and no doubt to its annual recurrence. 242 The scene on the altar 
V 1584, from Poetovio III, where the raven seems to be seizing some 
of the meat being roasted prior to the banquet of Mithras and Sol in 
order to carry it up to heaven, suggests that the bull's death was also 
understood as an offering to the gods; and perhaps that humans have 
to thank them also for its annual reiteration in the form of the renewal 
of life (PI. 19). Reverence for the gods and personal effort were key 
virtues demanded of Mithras' adherents in the West, just as they were 
in Iran, even though of course the western worshippers could have no 
inkling of the nature of the actual Iranian cult of Mithra, which was 
celebrated in an entirely different cultural context. Our sources often 
insist on the Persian origins of Mithraism: we cannot tell whether that 
claim was indeed true, or whether the main value of such an insistence 
was to be able to claim the prestige clinging to great antiquity. At any 
rate, the Mithraists believed it. 243 

Given the frequent ancient confusion between magi and Chaldaeans, 
between Persian priests and astrologers, it is of some interest that the 
Mithraic narrative evidently stressed certain astronomic-astrological con- 
nections. When a moment, or a series of moments, from the narrative 
was reduced to the standard Graeco-Roman iconographic form, the 
absence of an authoritative text expounding a normative interpreta- 
tion of images that were not subtended by well-known Classical stories 



241 As Insler 1978 suggested. The iconography leaves no room for doubt that Mithras 
hunts the bull down (Gordon 1988, 65), while the bull's s death only acquires its full 
meaning when understood as a sacrifice. I see no point in trying to force an opposition 
between hunt and sacrifice. 

212 Cf. Gnoli 1979, 740: "la simbologia della fecondita non e che uno dei modi in 
cui pud essere percepita l'alternanza vita-morte e in cui pud esprimersi la vicenda del 
dio mistico e dell' anima divina: la simbologia astrale e l'alternanza luce-tenebre ne 
constituiscono un' altra, non meno valida, modalita". 

2I! Cf. Statius, Theb. 1.7 19f; Origen, Contra Cels. 6.22; Porphyry, De antro 6; De abstin. 
4.16.2-3. 



102 CHAPTER TWO 

encouraged their interpretation in astronomical terms, among others. 
The complexity of the ancient astrological understanding of the heav- 
ens, with its risings and settings, houses and exaltations, dodecatamories 
and tropic degrees, provided ample material for ingenious minds. A 
dialectic that has recurred in the modern interpretation of the Mithraic 
icon may thus have established itself first in antiquity, and for much the 
same reasons. Many apparently contradictory or discrepant claims can 
be understood in this way. I think that Roman Mithraism responded 
positively to the demand to provide a coherent system of beliefs that 
gave an account of the relation between the stars, particularly the seven 
planets and the twelve signs of the zodiac, and human life, and so of 
the cosmic order as a whole, with this world at its centre. This became 
the basis of a cult that, as we shall see, paid particular attention to the 
outer frame of the reality or vision proposed by it, which I have sug- 
gested must be a key criterion of any proper religious system. 

It will by now be clear that the case of Mithras was different from 
the other mystery gods in that he had no personal experience of 
death himself. 244 Of course there was no imperative here. Neither Isis 
nor Cybele herself experiences death: they do so vicariously through 
their paredroi, Osiris and Attis. One way of increasing the parallelism 
between the three of them is Alfred Loisy's suggestion that the bull in 
Mithraism is an allegory of the god himself (Loisy 1930, 200). Under- 
standably, almost no one has followed him in this, though the idea has 
never been specifically rebutted. 245 



244 This fact creates a certain amount of difficulty concerning the most appropriate 
terms to use, as is clear from the Proceedings of Bianchi's Soteriology conference, for 
example in Turcan's point 2 in his summary (Bianchi and Vermaseren 1982, xvii), and 
in Bianchi's concession at the Seduta del 26 settembre: "et, un peu deplace, Mithra" 
(p. 883, cf. 885). Burkert 1987, 8; 41f; 84 repeatedly emphasises how untypical the 
cult of Mithras is in the context of Greek mystery cults. 

245 I have found the idea repeated only in the rather uninteresting — indeed super- 
erogatory — book by E. Wynne-Tyson, Mithras. The Fellow in the Cap (London 1952, repr. 
1972) 136, who borrows the phrase "the bull of Mithras killed by the god who was 
himself" from A. Weigall [1880—1934], The Paganism in our Christianity (London 1928). 
Robert Turcan has on several occasions criticised Loisy's 'exaggeration', e.g. Turcan 
1982b, 178 n. 50, and again in the same volume (Bianchi and Vermaseren 1982, 886); 
Beck ibid. 887 shares his view, but offers a considered reason for his rejection. Turcan's 
position is in large measure based on a no-nonsense reading of the imagery, but it is 
obvious that this need by no means coincide with the reality of the myth. The latter 
must have been far richer than allowed for by his remarks (e.g. 1981a, 4 If) a propos 
Cumont's belief that Mithras suffered because he was ordered to perform his heroic 
deed (e.g. Cumont 1929, 26: "condamne a creer le monde dans la douleur"). Here he 
emphasises the god's 'visage serein et decide': there is no sign of the regret of the sort 



SYSTEMS OF BELIEF 103 

It seems to me that insufficient attention has been paid to the relation 
between Mithras and his sacrificial victim, which is quite different from 
that between sacrificant and animal in civic cult. 246 The most important 
point is that, quite apart from the fact that it is a god who is sacrificing, 
the death of the bull gives rise to, or produces, life. This is the mystical 
paradox whereby death is no true death, but life. It is in this sense that, 
without pressing the point to complete identification, we can postulate 
a sort of bond or projection between Mithras and his victim. 247 

It is important to remember that the death of a primordial bull was 
a feature of several variants of the Zoroastrian and indeed pre-Zoroas- 
trian creation-myths. The usual Zoroastrian version was that Ahriman, 
the spirit of negation, had killed the cosmic bull. But recent research 



we might expect if there were a relation between sacrificant and victim. Against that, 
we might ask: what conclusion are we to draw if we note the direction of the gaze of 
the Father and the Holy Spirit in the presence of the crucifixion? Grimaces of pain 
are appropriate to mortals, as in the case of images of the Passion, but not gods. 

In his remarks in Bianchi and Vermaseren 1982, 899, Turcan urges that Mithras' 
epithet invictus could not be applied to Attis or Osiris, because they experience icd9r|. 
But in fact the character of a deus invictus is not incompatible with suffering or even 
with death, which is such an arcanum in religious thought — think of the paradox of a 
god, omnipotent, redeemer, and of course immortal, who is mastered and nailed to a 
cross. Here Jan Bergman's remarks a propos Osiris, who can be called invictus despite 
his having been torn to pieces (magnus deus denique summus parens invictus: Apuleius, Met. 
1 1.27.2), are very relevant (ap. Bianchi and Vermaseren 1982, 902f), as are those of M. 
Simon on Hercules (ibid. 903). For invictus used of Serapis, see RICIS 501/0146 and 
501/0147 (Rome, first quarter IIP); 603/1 102 (Tarraconensis, same period) 616/0205 
and 0207 (Sarmizegetusa, both by procuratores Aug., same period); 617/0101 (Moesia 
Sup., same period); cf H. Engelmann, The Delian Aretalogy of Sarapis. EPROER 44 
(Leyden 1975) 24; Hornbostel 1973, 273f (coins and Kleinkunst). It seems likely that 
it was a borrowing from the cult of Mithras at the time of Caracalla: Malaise 1972a, 
1 94f and 44 1 . It must however be admitted that Attis is never called invictus. 

246 This is one of the best points in Eisner 1 995, 1 90-245, cf. 1 998, 205-2 1 . Mithras is 
not the recipient of the sacrifice but himself the sacrificant; consequently the tauroctony 
is not a representation of an actual sacrifice but a symbolic image. This is precisely 
one of the great innovations of Mithraic religious art, in that the image in context 
annuls the distance between reality and and its representation, the distance taken for 
granted in the case of images of sacrifice in civic religion. That is, in the mithraeum, 
the cult-image and the ritual are united in a single physical and temporal space, in 
an allegorical world whose referents are to be found the 'other world'. As Eisner says 
(1995, 218—21) Mithraism required a strong commitment and a complex ritualisation 
in order to resolve the contradictions within such a symbolism. In the later book (1998), 
intended for a wider audience, Eisner just presents his previous conclusions without 
the decent subtleties of the earlier book. 

2 " J. Bergman suggested at the Soteriology conference that 'Mithras and the bull' can 
be considered a couple parallel to Cybele and Attis, Isis and Osiris, each understood 
as two persons, or, better, roles: "Le Taureau meurt, mais Mithras reste invictus" (ap. 
Bianchi and Vermaseren 1982, 902). 



1 04 CHAPTER TWO 

has produced some (rather speculative) evidence that there may have 
been a very ancient pre-Zoroastrian Iranian version in which Mithras 
sacrificed the bull in order to free light, and the life-principle, from the 
confines of darkness (Kreyenbroek 1994). It may be that victim and 
slaughterer, created by the same divinity, retained some trace of their 
common parentage. Whatever the actual process, it seems certain that 
for adherents of the cult in the second century AD the tauroctony had 
a value different from that of earlier times, even if there was some 
continuity. And in the Roman reworking of the cult, it is possible that 
a connection was established between Mithras and the bull such that 
the god's essential nature could be understood as similar to that of the 
other analogous mystery deities. 248 

In fact it seems probable that it is the bull's blood that produces 
life and so salvation, since it evidently contained the animal's divine 
essence. 249 The bull thus contains the potentiality that the other systems 
we have looked at assign to the god (Osiris, Attis). In Mithraism the 
god himself mediates between the vessel that contains the potential for 
life and creation itself. Thanks to the sacrifice Mithras performs, the 
bull becomes part of him, in the sense that if it were not present, there 
could be no demiurgic act. It is therefore quite unnecessary actually 
to identify the bull with Mithras, as Loisy did; it would be enough for 
them, inseparable as they were, to have been considered interchange- 
able as agents of the effects of the sacrifice. Indeed, Porphyry seems to 
suggest as much atDe antro 18 (= Sanzi 2003, 421: Mithras 8.5), where, 
in the context of the astrological connection of the Moon with Taurus 
and the well-known origin of bees [= souls] from dead bulls, he says: 
koc1 \|A)xai 8' ^iq yeveoiv kvoaca povyevefi;, koc1 ^ovxXonoc, 9eoq 6 xf|V 
yeveow A,eA,r|96xK)q f aKovcov f, 'and souls come into genesis born of 



248 Casadio has recently suggested that the scorpion is attempting to castrate the 
bull, which would build a bridge between bull and Attis (2003, 266f.). 

249 I am perfectly aware of the difficulties involved in equating the sacrificial victim, 
who must of course be a sacred object in some sense, with the god himself to whom 
the offering is made. As H. Hubert and M. Mauss put it, "a divine victim is not a 
victim-god": Sacrifice: Its Mature and Function. Tr. W.D. Halls (Chicago 1964) 78 (= Essai 
sur la nature et fonction du sacrifice, Annie sociologique 2 (1898) 29—138 at 1 15f). In their 
view, sacrifice may exalt the victims in such a way as to divinise them directly, but in 
order for a god to be a victim there must be an affinity between his nature and that of 
the victims. Agrarian cults typically fulfil this condition. Moreover, a sacrificial victim 
need not be divine, but once divine victims become thinkable, we cannot exclude the 
possibility that the Mithaic bull might have been one. 



SYSTEMS OF BELIEF 1 05 

bulls, and the Bull-stealing God is he who [Psteals; ?brings; ?promotes] 
genesis secretly'. 

Other allegorical interpretations might be invoked here, though they 
are of course merely speculative. C.G.Jung, for example, suggested the 
tauroctony might represent the benefit given mankind by the victory 
of the spiritual, represented by Mithras, over animality symbolised by 
the horned bull. 250 Disregarding the moral judgement implied there, 
the god might represent the civilising principle by contrast with primal 
nothingness: reality does not yet exist but is merely potential, confined 
inside an entity of extraordinary vital force represented by the prime- 
val bull taking its rest in a sacral space (the temple depicted on some 
monuments) out of time. 

However that may be, we can surely say that Mithras' task was to 
free the forces contained within the bull and give life the chance to 
come into the world. His destiny is fulfilled in the killing: gods and 
mankind acknowledge the heroic exploit that produces the generation 
of the cosmos. 251 If the god's birth implies his control over the cosmic 
order, the tauroctony is an extraordinary allegory of universal scope 
that was also reproduced in initiation. According to Origen's quotation 
of the account by the anti-Christian philosopher Celsus (written c. AD 
1 7 7-80), some Mithraists at any rate imagined that the soul ascended 
through the seven planetary orbits, and passed from there into the 
sphere of the fixed stars, f) anXav^c, 7iepio8o(;. 252 Initiation took place of 
course in the mithraeum, which we know was an image of the cosmos. 
Tauroctony, initiation and mithraeum frame the imaginary universe 
within which Mithras exercises his dominion over the cosmos. 233 It is 
inside the mithraeum that real and mythical time interpenetrate such 



230 L'homme et ses symboles (Paris 1964) 147. Of course, the bull, as a repository of 
cosmogonic power, could never represent either chaos or animality, but simply contains 
life. It is for that reason it is sought out by the demiurge and sacrificed, thus liberating 
the forces of nature that give rise to life. 

2,1 Cf. Zwirn 1989, 9. This may be one of the significances of the depiction on some 
German and Danubian complex reliefs, above the tauroctony, of the assembly of the 
(twelve) gods, representing heaven as the abode of the gods, e.g. V 1292, Top no. 1 
(PI. 17 here) = C.R. Long, The Twelve Gods of Greece and Rome. EPROER 107 (Leyden 
1987) 27f. no. 1 (unhelpful commentary on p. 288); 1128 = Schwertheim 1974, 84f. 
no. 611 (Danubian, found at Nida/Heddernheim III) scene 1 1; V 1475 = Selem 1980, 
84f. no. 15 with pi. XVII. 2 (Siscia, Pannonia Sup.). 

252 Contra Gels. 6.22 = Sanzi 2003, 416: Mithras no. 7; cf. Turcan 2000, 11 Of; sug- 
gesting that this represents a 'Great Sideral Week', not a posthumous soul-journey I 
incline to prefer Beck's notion of a true soul-journey (e.g. Beck 2004a, 68—75; 267f). 

253 Gordon 1988; Beck 2006, 102-12. 



1 06 CHAPTER TWO 

that the cosmogonic process can in a sense continue there; where the 
grand machinery whirrs that keeps the universe going, with its heavenly 
gates for the souls to enter the world and then leave it after death. 

I would then want to suggest a certain parellelism between Mithraism 
and the other cases we have already examined: Mithras is a god who 
controls the heavens; by virtue of his own power, he bestows order on 
nature, brings about agricultural and sexual fertility, and provides his 
followers with a example to follow. Mithras does indeed not himself 
suffer. But we might think of the bull as a sort of substitute figure, not 
of course a manifestation of the god himself, but as a being whose 
nature is divine, all the more so because he becomes the victim whose 
death is the necessary precondition of life. 254 

2. Humankind in the World 

Dragonflies drift on the river, 

Their faces look upon the face of the Sun, 

(But then) suddenly there is nothing. 

Gilgamesh Xl.vi = Dalley 1989, 109 

II est tres rare qu'un homme peut supporter sa 
condition comme homme . . . 

Andre Malraux, La condition humaine (1933) 

The poverty and vagueness of the sources available make it impossible 
to write a proper anthropology of the mysteries. It is difficult to engage 
with the problem, since there has been little exploration even where the 
opportunity offered. 255 Nevertheless it seems worth offering a preliminary 
sketch here. The absence of theories developed by the adherents of the 



2,1 Cf. Turcan 2000, 76. I might here mention the rather simplistic account — he 
calls it 'holistic' — of P. Bilde, The Meaning of Roman Mithraism, in J. Podemann 
Sorenson 1989, 31-47. 

255 A study such as I suggest might have found a place in the volume edited by 
J. Bleecker, Anthropologic religieuse. L'homme et sa destine a la lumiere de I'histoire des religions. 
Studies in the History of Religions 2 (Leyden 1955). Among other articles there, let 
me draw attention to that of R. Pettazzoni, La condition humaine, pp. Iff, where I 
found my epigraph from Malraux. He recalls the seven sages of Greek tradition, who 
exhorted human-beings to confine themselves to their human condition, know their 
own limitations, abstain from excess and think and act like mortals. The themes of 
J. Ries (ed.), Traite d' anthropologie du Sacre,l: Les origines et le probleme de I'Homo Religiosus 
(Paris 1992), have no direct relation to my aims here. R. Lapointe, Socio-anthropologie du 
religieux, 1: La religion popuhire au peril de la modernite (Geneva 1988) and ~ 2: Le eerele enchante 
du croyance (Geneva 1988) is an interesting general introduction to popular religion. 



SYSTEMS OF BELIEF 107 

mysteries makes it still more difficult to characterise them in terms of 
their anthropology. We lack information about their view of the world, 
and the relation between humans and their environment. 256 We can at 
most discuss the influence that different philosophical traditions may 
have had on the authors who transmit information about the mysteries, 
but of course that is a far cry from the mysteries' own conception of 
the human condition. It is likely that some leading ideas later picked 
up by Christianities had already found a place in the thinking of the 
adherents of the mysteries. These may include the idea that the body 
is the prison of the soul (assuming the existence in the mysteries of an 
opposition of this kind) or, more plausibly, the metaphor of death as a 
liberation, which does seem to have played a part in initiation into the 
mysteries and is often referred to by Paul. 237 As for the Christianities, 
they offered answers appropriate to every type of issue that was raised. 
The mysteries could not offer a formula remotely comparable to the 
thought that God has created human-beings so that they may rever- 
ence and worship him for ever. 208 In fact, such a formulation drives a 
wedge between the object of human life and that of the remainder of 
creation, since the latter has no other role than to facilitate or serve the 
reproduction of humanity so that it may devote itself to the veneration 
of God. 259 Human-beings were thus placed squarely at the centre of 



2: "' However I fully agree with Turcan's formulation: "All these cults that originated 
in the east, although very hellenised, provided their initiates with an explanation of the 
world and of man in the world. Their cosmic perspective coincided with the social and 
supra-national reality of an universal empire such as that of Rome": Cultes mysteriques 
et culture classique dans le monde romain, in A. Caquot and P. Canivet (eds.), Ritualisme 
et vie interieure. Religion et culture. Collogues de 1985 et 1987. Societe E. Renan: Histoire des 
Religions. Le Point theologique 52 (Paris 1989) 152-68 at 155. 

2: " Death is the point of entry into an unfathomable darkness from which humans 
must be saved and from which they can be redeemed by divine power. As such, death 
becomes the key to an incredibly rich allegorical world: D.E. Aune, Human Nature and 
Ethics in Hellenistic Philosophical Traditions and Paul: Some Issues and Problems, in 
T. Engberg-Pedersen, Paul in his Hellenistic Context (Edinburgh 1994) 291—312. 

258 This is the burden of Athenagoras, De resurr. carnis, written under Marcus Aure- 
lius. He concludes (22): "And we shall make no mistake in saying that the aim of a 
an intelligent life and rational judgement is to be occupied uninterruptedly with those 
objects to which the natural reason is chiefly and primarily adapted, and to rejoice 
unceasingly in the contemplation of Him who is, and of his decrees, notwithstanding 
that the majority of men, because they are affected too passionately and too violently 
by things below, pass through life without attaining this object" (tr. H.J. Richardson). 
Cf also Lactantius, De ira 14.1—2. 

2,9 Socrates is made to say as much at Xenophon, Mem. 4.3.9f. Euthydemus replies: 
6pS yap [xa i^tpa . . .] oikax; vrcoxeipia yiyvonEva xoti; av9pco7ioi<;, coaxe xpfiaGcu at>xoI<; 



1 08 CHAPTER TWO 



creation, at the fulcrum of the world order, such that nothing had a 
meaning except in relation to humanity. 260 

From this point of view, animals and plants had no other purpose 
than biological reproduction, whereas human-beings had a further 
dimension, the spiritual, which their culture gave them in the form of 
a particular religion. Under these circumstances Stoicism and other 
hegemonic philosophies of antiquity simply could not raise the issue of 
the rights of animals. Nor could the suffering of animals be considered 
a sufficient reason to develop a positive theory of such rights. 261 The 
problem was only an issue in relation to the claims of vegetarians. 
These acquired a special relevance in the context of religious debates 
over blood-sacrifice, which were of some importance in the Empire 
at the time of the floruit of the mysteries. 262 This is the context of 
Porphyry's workD<? abstinentia, which was written around AD 270 under 
the influence of Plotinus' views, 263 and is the most important theoretical 
discussion of the relation between humans and animals to have come 
down to us from antiquity 264 There he claims, apparently following 
an argument not recorded by Plutarch in De sollertia, that animals and 
plants have different natures, since animals have sensation, can suffer, 
be afraid, suffer loss; and it is therefore meaningful to say that it is 
possible for them to be maltreated. The same cannot be said about 



o xi av PovXtovToa, "I see that animals . . . are born so subordinate to men that the latter 
can make use of them as they like". 

21 '° Cyprian's position was different, however, cf. J. P. Burns, Cyprian's Eschatology: 
Explaining Divine Purpose, in A.J. Malherbe, EW. Norris and J.W. Thompson (eds.), 
The Early Church in Context: Essays in Honour of E. Ferguson. Suppl. Novum Testamentum 
90 (Leyden 1998) 59—73. For an introduction to these problems in the main religions 
today, cf. J. Holm (ed.), Human Nature and Destiny (London and New York 1994), esp. 
the contribution on Christianity by D. Davies, pp. 39-70. 

261 The idea of the superiority of animals was quite dead at latest by the fifth century 
BC, except as a trope to underscore the imperfections of humanity, cf. A.O. Lovejoy and 
G. Boas, Primitivism and Related ideas in Antiquity (New York 1965 [1935']) 389-420. 

21)2 J. Haussleiter, Der Vegetarismus in der Antike (Berlin 1935); D. Tsekourakis, Pythagore- 
anism or Platonism and Ancient Medicine? The Reasons for Vegetarianism in Plutarch's 
Moralia,ANRWII.36.l (1987) 366-91: Saelid Gilhus 2006, 64-77. C. Osborne, Ancient 
Vegetarianism, in J. Wilkins, D. Harvey and M. Dobson, Food in Antiquity (Exeter 1995) 
214—24 is simply a commentary on Porphyry; she completely disregards other sources 
and modern studies (even Haussleiter). 

263 Cf. J. Bidez, Vie de Porphyre (Ghent 1913); J. Bouffartigue and M. Patillon, Porphyre: 
De I'abstinence 1 (Paris 1977) xviii-xix [Bude]. 

264 See the excellent translation and commentary by G. Clark, Porphyry: On Abstinence 
from Killing Animals (London 2000); also D.A. Dombrowski, Porphyry and Vegetarianism: 

A Contemporary Philosophical Approach, AMRWIIM.2 (1987) 774-91. 



SYSTEMS OF BELIEF 1 09 

plants {De abst. 3.19.2). 265 My guess is that the mysteries contributed 
to installing normative anthropocentrism into the sacrificial system of 
antiquity The emperor Julian, who knew a good deal about the mys- 
teries and attempted to restore animal-sacrifice to its earlier centrality 
seems to suggest as much. Two centuries earlier, Apuleius, an expert 
on the same area, set out to imagine the sufferings of a human-being 
who takes on an animal's body. But his account is essentially no dif- 
ferent from Julian's: he just confirms the anthropocentric view of his 
contemporaries. We have to wait until the twelfth century to find Moses 
Maimonides of Cordoba claiming that to inflict suffering on animals 
corrupts our characters and makes us vicious. 266 The dominant view in 
classical antiquity was that human dominion over all other creatures is 
an incontrovertible fact. An illustration might be Ovid's description of 
the creation of man, where he stresses that, whereas the other animals 
were forced to look down at the ground, humans were endowed with 
a face that enabled them to gaze up at heaven and the stars: 



263 Cf. R. Sorabji, Animal Minds and Human Morals. The Origins of the Western Debate 
(London 1993) 208. 

266 Guide 3.17 and 48. The first passage is part of a discussion of whether divine 
providence also includes animals; we ought not to make them suffer unnecessarily, 
only when our needs require it, and we ought not to kill them out of cruelty or whim. 
Maimonides' thinking is clearly marked by the tension between his conception of per- 
sonal freedom (perhaps exaggerated by modern scholars) and the destiny prescribed 
by God: M. Sokol, Maimonides on Freedom of the Will and Moral Responsibility, 
HThR 91 (1998) 25—39. The bibliography on Maimonides' ethics is immense. For the 
passages themselves, see, apart from the excellent annotated Spanish translation by 
D. Gonzalo Maeso, Rabbi Mose ben Maimon (Maimonides): Guia de perplejos (Madrid 1984), 
the abridged English ed. by J. Guttmann (tr. C. Rabin), The Guide for the Perplexed (London 
1952 and often repr.). I have not seen the new Kegan Paul edition: Guide for the Perplexed 
(London 2006). R.L. Weiss, Maimonides' Ethics: The Encounter of Philosophic and Religious 
Morality (Chicago and London 1991) 167f£ outlines Maimonides' concept of compas- 
sion in relation to his respect for animals, cf. S. Rosenberg, La etica en Maimonides, in 
J. Pelaez (ed.), Sobre la viday obra de Maimonides (Cordoba 1991) 455-62; A.J. Heschel, 
Maimonides (Barcelona 1984). On Leo Strauss' interpretation of Maimonides (with 
S. Pines, Guide for the Perplexed, 2 vols. [Chicago 1974]), seeJ.A. Buijs, The Philosophi- 
cal Character of Maimonides' Guide — a Critique of Strauss' Interpretation, in idem 
(ed.), Maimonides: A Collection of Critical Essays (Notre Dame, IN 1988) 57-71. Finally, 
A. Sanders, Dear Maimonides. A Discourse on Religion and Science (Northvale, NJ and Lon- 
don 1996) is an unpretentious attempt to reflect on this issue from the Jewish point of 
view, underscoring the paradoxes of the culture. By means of the device of a series of 
letters between Maimonides and his favourite disciple Yosef ben Yehudah ibn Shim'on 
(to whom the Guide was dedicated), he entertainingly discusses the main historical and 
scientific issues from the twelfth and the twentieth centuries, highlighting how much 
Maimonides simply did not know. 



110 CHAPTER TWO 

Pronaque cum spectent animalia cetera terrain, 
os homini sublime dedit caelumque tueri 
iussitet erectos ad sidera tollere vultus. 267 

Though voices were occasionally raised against this representation of 
the matter, they were ignored. That the mysteries went along with the 
dominant view is undeniable, even though, given the special features 
of Egyptian religion, viz. the zoomorphism of its gods, the adherents 
of the Egyptian cults might have been an exception. 268 Nevertheless 
we have explicit evidence in an anecdote in Aelian's De natura animalium 
1 1.34f. that the worshippers of Serapis might consider animals as tools 
in the hands of gods made to serve the needs of mortals. 269 No doubt 
the adherents of the other mysteries thought just the same. On the other 
hand, an analogous belief in the natural inequality of human-beings 
can be found in Hellenistic-Roman thinking, despite various pleas that 
all humans are essentially equal. 2 ' Ideas about the equality or inequal- 
ity of humans are tricky to talk about, since one cannot generalise to 



267 Ovid, Met. 1 .84—86 with the commentary by F. Bomer. On the context of Ovid's 
thought, see F. Lammli, Vom Chaos zum Kosmos: Zjir Geschichte einer Idee (Basel 1962) 9ff 
This theme of where one looks seems to have remained important, since among the 
animals forbidden to those who celebrate the Phrygian festivals was the pig, whose 
flesh, according to Julian, Or. 5. 1 77bc, is thought suited to the gods of the underworld 
but not allowed on the tables of the worshippers of Cybele because of its shape and 
mode of life, and the character of its meat; moreover it does not raise its eyes to 
heaven, not only because it does not want to do so but because it is so made that it 
cannot lift its gaze. 

268 We have virtually no information about the implications of this for the worshippers 
of Isis in the Roman period. But there is surely a hint in Diodorus Siculus' eye-witness 
report (1.83.8) that a Roman in Egypt paid with his life in 59 BC for involuntarily 
killing a cat, so seriously might respect for sacred animals be taken there, cf. K.A.D. 
Smelik, The Cult of the Ibis in the Graeco-Roman Period, in Vermaseren 1979, 
225—43. The later Christian claim that they were zoomorphic idolaters needs to be 
related not just to the obvious iconography well-known to contemporaries but also to 
alimentary practices and concomitant ideas it is hard to get a purchase on. Some of 
these Christian texts, such as Prudentius, Apotheos. 195f; In Symmach. 2. 3542! and 532 
(both on Anubis), and the ridicule of Egyptian gods and food at Peristeph. 10.256-60 
(a notoriously unreliable text) will be found in J.-C. Grenier, Anubis alexandrin et romain. 
EPROER 57 (Leyden 1977) 69—83, where he discusses the implications of zoomorphism 
for followers of Isis; cf. too Witt 1966, 136. 

269 The story concerns a young man who, being seriously ill, dreamed Serapis told 
him to buy a live moray eel and plunge his hand into the creature's tank. When he 
did so, the eel bit him, but when it was pulled oif, it was found it had also 'pulled off 
the sickness: xmr\ps.xiq, . . . Geov Geparceiai; r| uttpoava. 

270 J. Chesneaux, Egalitarian and Utopian Traditions in the East, Diogenes 62 (1968) 
76-102; C. Mosse, Les utopies egalitaires a l'epoque hellenistique, Revue Historique, 
241 (1969) 297-308; J. Ferguson, Utopias of the Classical World (London and Ithaca, 
NY 1975); D. Dawson, Cities of the Gods: Communist Utopias in Greek Thought (New York 



SYSTEMS OF BELIEF 111 

other areas from individual statements about, say, social, moral, political, 
economic or religious equality. 271 For example, the famous utterance of 
Paul: onov otjk evi "EA.At|v kocI 'IooSouoc;, 7iepixour| kcxi aKpopvcmoc, 
pappapoq, Zic69r|<;, Sov>A,o<;, eJieviBepoc;, ocAAoc navxa Kai ev naaiv 
Xpioxot;, "Where there is neither Greek nor Jew, circumcision nor 
uncircumcision, Barbarian, Scythian, bond nor free: but Christ is all in 
all" (Coloss. 3.11), is entirely restricted to this last context. 2 ' 2 Moreover, 
although it has often been taken as a declaration of general equality 
within Christianity, it in fact refers to a purely theoretical access to the 
godhead, which the Church itself busily restricted over the following 
centuries of its history. 

The very character of the mysteries, as well as the fact that they were 
embedded in the religious universe of the Roman Empire, means that 
it is difficult to consider them as egalitarian systems. Conditions were 
not favourable to the development of egalitarian Utopias, and, to the 
degree that they sought social acceptability, as expressed in their abil- 
ity to attract adherents throughout the Empire, they abandoned any 
pretensions they might ever have entertained to further socio-political 
change. 273 They had undergone a good deal of opposition before being 
able to enjoy the tolerance that turned them into effective instruments of 
social integration under the Roman oligarchy. This process is still more 
pronounced in the case of mainstream Christianity, since its quietistic 
marriage to the institutions of the state moored the notional equality 
of mankind in the intangible universe of the axiomatic, leaving the 
actual assertion of the claim to others. 

Turning the mysteries again, we find that they asserted a fundamental 
distinction between those who had been initiated and those who had 



1992); M. Schofield, Saving the City: Philosopher-Kings and Other Classical Paradigms (London 
and New York 1999) 31-68. 

2/1 Cf. G.J.D. Aalders, Ideas about Human Equality and Inequality in the Roman 
Empire: Plutarch and some of his Contemporaries, in I. Kajanto (ed.), Equality and 
Inequality of Man in Ancient Thought (Helsinki 1984) 55-71 at 56. To the contrary of 
much that has been written on the topic, Christianity initially changed little here, cf. 
H. Giilzow, Christentum und Sklaverei in den ersten drei Jahrhunderten. (Bonn 1969); P.D.A. 
Garnsey, Ideas of Slavery from Aristotle to Augustine (Cambridge 1996) 189-235. 

2/2 In the parallel passage at Galatians 3.28, he adds the opposition between men 
and women. 

273 Note the remarks of Walters 2000, on the status and interests of the wealthy 
women at Athens during I-II p whose families erected funerary statues of them in Isiac 
dress, all with knots, some with sistra (cf. eadem 1988). She compares them to the 
procession of initiates at Apuleius, Met. 11.10. 



112 CHAPTER TWO 

not. One category of persons was thus granted superiority, at least of 
a moral kind, over another, as we see in a passage cited by Stobaeus 
and generally believed to be from Plutarch's De anima: 

Thus we say the soul that has passed thither is dead (6A,coA,evoa), hav- 
ing regard to its complete (eic, to oA,ov) change and conversion. In this 
world it is without knowledge, except when it is already at the point of 
death; but when that time comes, it has an experience like that of men 
who are undergoing initiation into great mysteries (oi xtktxaiq \i.tj6Xa\q 
Koaopyia^oiievoi); and so the verbs xeXevxav (die) and teAxiaSou (receive 
initiation), and the actions they denote, have a similarity. In the beginning 
there is straying and wandering, the weariness of running this way and 
that, and nervous journeys through darkness that reach no goal, and then 
immediately before the consummation every possible terror, shivering and 
trembling and sweating and amazement. But after this a marvellous light 
meets the wanderer, and open country and meadow-lands welcome him; 
and in that place there are voices and dancing and the solemn majesty 
of sacred music and holy visions. And amidst these he walks at large in 
new freedom, now perfect and fully initiated, celebrating the sacred rites, 
a garland upon his head, and converses with pure and holy men; he 
surveys the uninitiated, unpurified mob here on earth, the mob of living 
men who, herded together in murk and deep mire, trample one another 
down and in their fear of death cling to their ills, since they disbelieve 
in the blessings of the other world. 

Stobaeus, Eel. 4.52.48 = frg. 178 Sandbach 

On the other hand, within each cult, the array of responsibilities, duties 
and ranks or grades reproduced the social stratification of Roman 
society. Whatever the appeal to the ideal of equality expressed in the 
blanket terms: 'initiates', 'companions', 'brothers', 'co-religionists' and 
so on, the reality made all too clear that they had embraced the hier- 
archical option. The true function of these terms was to differentiate 
the members of these groups from those outside (Gordon 1972). Such 
a hierarchical structure reinforced the hegemonic image of man and 
the general anthropocentrism which was the real basis of their offer 
of personal salvation. Unfortunately we have no direct information 
about the mysteries' conception of the meaning of life or the role of 
mankind in the world. Although their central theme was individual 
salvation, the promise of a blessed life in the world beyond went hand 
in hand with a sort of mystic union with the divinity, even if we are 
not in a position to fill in any of the details. Yet the myths may give 
us one or two hints. 

I would argue that the myths do in fact give us a more or less explicit 
account of the place of man in the world. Of course, just as in other 



SYSTEMS OF BELIEF 113 

contexts, we cannot expect to get straight answers to straight questions. 
The myths do not tell us about existential anxieties, but they can give 
us some insight into the more or less specific solutions offered to them. 
The narratives imply a behavioural norm that is paradigmatic for the 
adherents of the cult. In this sense, the human situation in the world 
does relate to the system of beliefs, articulating a reality separate from 
that of the model of ethical behaviour that the myth also contains. We 
shall look at this more closely in the following chapter on the system of 
values. The myth of Attis and Cybele for example is a good illustration 
of the second level of analysis, the relation of man to his environment, 
since it is clearly concerned with the annual cycle of nature as repre- 
sented by its apparent death in winter and resuscitation in spring. That 
the myth reproduces the agricultural cycle has been clear since Frazer 
(even if later work has discovered additional readings or even repressed 
the point), but there is no need to infer that the myth could only have 
been created after the development of agriculture, since hunter-gather- 
ers were equally capable of noting the regularity of natural cycles and 
creating images capable of enduring in the collective imagination. 

Quite apart from this, we must of course remember that the myths 
have themselves evolved: the original narrative has been altered in order 
to adapt them to changes in the relations of production. Although 
we can of course hardly hope to gain access to the earlier versions, 
the nucleus of the story persists through time. At the same time, the 
variants may indeed succeed in influencing the central mythologem: 
the grip of history upon the narrative must prejudice any hope of 
using the latter as a key to some claimed universal reality. The idea of 
'myth in motion' is important because of its implications for the sense 
in which we can say we know a myth. In the cultural context we are 
concerned with, the myth is clearly moving away from a set of images 
developed — insofar as we can reconstruct them — in the context of 
agrarian-pastoral societies. 274 We may therefore gain the impression that 
such myths, in the form they have come down to us, could only have 
developed in the context of historical formations which have devel- 
oped this type of relations of production. The important thing, then, 
is to emphasise how agricultural production is appropriated into the 
myth-cycle and revealed to be subject to the divine order rather than 



274 [The Spanish adj. is agropecuario, which deserves to be domesticated into English. 
TV.] 



114 CHAPTER TWO 

to the whim of chaos. This way of presenting things tends in turn to 
encourage the religious subordination of the primary producer to the 
established order, itself legitimated by the divine will. 

At the same time this order is itself organised in the interests of those 
who have social and ideological control over the political formation. 
Belief in a god who controls nature and regulates its productivity 
fits neatly with the politico-social domination of the class that is the 
primary beneficiary of that productivity. It was this that encouraged 
acceptance of the mystery divinities, while other traditional gods were 
'orientalised': the goddess Ops for example, curiously enough, was 
linked to both Cybele and Isis. 273 

Evidence that has no direct connection with the myths allows us up 
to a point to see how the deity intervenes to establish order, thus pro- 
viding an epistemologically satisfactory image of how the world really 
works. Nancy Shumate has offered a reading of Apuleius' Metamorphoses 
that shows how the world-view of the protagonist-narrator crumbles as 
he comes into contact with certain 'realities' that confound his rational 
account of things. The eredulitas of someone who feels himself secure 
in a cultural context is radically altered by the experience of alter- 
ity represented above all by magic, which undermines a social order 
founded on the opposition between the central and the sub-cultural. 
The challenge to his old beliefs destroys his fides and brings chaos into 
his world. The impossibility of understanding how the world indeed 
works, the cognitive bouleversement, turns the protagonist-narrator into 
a donkey, and in that shape he experiences the irrationality of a world 
beastly and confused from which he is only freed by Isis. 276 The goddess 
instructs him in a quite different epistemological order that is stable 
and clear, thanks to which Lucius is able to comprehend his place in 
the world. 27 ' Apuleius' novel can thus be understood as an allegory 
by means of which the author expresses his conviction that personal 



2 '"' Cf. Varro, LL 5.57: Principes del Caelum et Terra. Hi dei idem qui Aegypti Serapis et 
his, etsi Harpocrates digito significat, ut taceam. Idem principes in Latio Saturnus et Ops; with 
P. Pouthier, Ops et la conception divine de VAbondance dans la religion romaine jusqu'a la mart 
d'Auguste. BEFAR 242 (Rome 1981) 303-10. 

276 Cf. J.J. Winkler's phrase "a brutal odyssey through a strange land": 1985, 59. 

2 " Cf. Shumate 1996. This is a solid piece of work, and I welcome her attempt to 
integrate Book 1 1 with the rest of the novel, but I find her claim that Apuleius' conver- 
sion is purely cognitive or epistemic very reductive. I hope in the section on Isiac ethics 
(Chap. 3.3) to convince the reader that the change was moral as much as anything. 



SYSTEMS OF BELIEF 115 

identity can be independent of the public domain, a preliminary step 
to the defence of Isis' power to grant individual salvation. 278 

In the case of Mithras too the myth suggests that acceptance of a 
god capable of imposing order on the forces of nature may be instru- 
mental for man in the context of production. It is this that gives the 
impression that Mithraic initiation "seems to be interested not so much 
in preparing the soul for its existence after death as in fitting man for 
his place in this world, accepting it not as better but as the only possible 
one, because it is guaranteed by the sacrificial act of the founding god, 
whether demiurge or hero" (Chirassi Colombo 1982, 317). 

No less important is a topic that has already cropped up in other 
contexts, but which needs to be considered on its own account, since 
it concerns mastery of the contingent not in relation to the after-life 
but to the changes and chances of the real world. The literature of 
the Principate is full of references to the profound pre-occupation of 
contemporary society with the consequences of the uncontrollable forces 
of destiny. On the one hand, philosophers developed lines of argument 
to justify the given world, and offered strategies by means of which their 
followers could prepare themselves against reverses of fortune, should 
they come, and accept their entire account of the nature of things. On 



2/8 Cf. Y.L. Too, Statues, mirrors, gods: controlling images in Apuleius, in J. Eisner 
(ed.), Art and Text in Roman Culture (Cambridge 1 996) 133-52 at 152. Note also Finkelpearl 
1998, 216: "There is an identification here of author and protagonist, not in terms of 
life experiences or religious kinship, but in terms of the parallel development of the 
events of Lucius' life and the movement of the novel". As I have already suggested in 
the Introduction, I accept many of the arguments of Reardon 1971, Hagg 1980 and 
Winkler 1985 regarding the over-all ironic tone of the novel. In the context of Met. 
Bk. 11, however, I see this position, with Anderson 1982, 84 (cf. idem 1984, 85), as a 
reaction to the excesses of the Kerenyi-Merkelbach allegorical approach to the genre as 
a whole (Kerenyi 1927; Merkelbach 1962) and to the credulousness of the older naive 
readings of Met. 1 1 as straight or 'sincere' documentation (e.g. Nock 1933a, 138—55; 
A.-J. Festugiere, Personal Religion among the Greeks [Berkeley 1954] 68-84; Dodds 1965, 3; 
P.O. Walsh, Lucius Madaurensis, Phoenix 22 [1968] 143-57; Griffiths 1975). We need 
a perspective on Bk. 1 1 that allows us to see it both as distanced and yet also as an 
imaginative (fictional) insight into a particular experiential world, cf. Mimbu Kilol 1994 
(though I think the main thesis, that the first ten books correspond to the prolegomena 
to initiation, very far-fetched). I find Methy 1999 both slight and misguided here: we 
can surely allow both that Apuleius was intelligible to his readers and that they could 
make sense of the book's religious world, just as we can, even if his intention was 
also to make fun of it, e.g.: Necfuit nox una vel quies aliqua visu deae monituque ieiuna (19); 
nee non et equum quoque ilium meum reducentes (i.e. Lucius' friends), quern diverse distractum 
notae dorsualis agnitione recuperaverant (20); et ecce post pauculum tempus inopinatis et usquequaque 
mirificis imperiis deum rursus interpellor et cogor tertiam quoque teletam sustinere ...et hercules iam 
de fide quoque eorum (i.e. the priests') opinari coeptabam sequius (29.1-3). 



116 CHAPTER TWO 

the other, institutionalised religion had always offered the hope that it 
might be possible, thanks to an entire gamut of divinatory means, to 
interpret the supposed will of heaven. Dreams of course were one of 
the main channels of information regarding the future and divine inten- 
tion (Hidalgo 1992). At the same time mechanisms were developed to 
control those specialists who claimed access to such knowledge and to 
be able in return for a small sum to state how things would turn out, 
and even how to avert such futures. 

The performance of miracles became a primary focus of religious 
attention. 279 Charismatic figures, so-called avSpec; BeTot, sought to 
trump one or other famous miracle, revealing the intense interest of 
the consumers of new religious experiences in the performance of such 
events. 280 Many hoped thereby to be able to escape from their wretched 
fates and of course were inclined to fall for the one whose offer was most 
attractive. Lucian railed at such behaviour, 281 but it is precisely thanks 
to his annoyance and the enthusiasm of others that we know a good 
deal about the performances of people like Alexander of Abonoutei- 
chos, Peregrinus Proteus, or Apollonius of Tyana just to name the most 
well-known. 282 Miracle became part of the sociology of knowledge, and 



2,9 Here one must refer to the excellent book by H. Remus, Pagan-Christian Conflict 
over Miracle in the Second Century (Philadelphia 1983), who deftly analyses the well-known 
cases and fits them lucidly into their social context. Note also D.L. Tiede, The Charis- 
matic Figure as Miracle-Worker (Missoula MO 1972): G. Filorama and S. Roda, Religione 
popolare e Impero Romano, Studi Storici 23 (1982) 101-18 at 110; G. Anderson, Sage, 
Saint and Sophist. Holy Men and their Associates in the Roman Empire (London and New 
York 1994); J.A. Francis, Subversive Virtue. Asceticism and Authority in the second-century Pagan 
World (University Park, PA 1995). Peter Brown's late-antique Syrian holy-men thus 
had a very long lineage: P.R.L. Brown, The Rise and Function of the Holy Man in 
Late Antiquity, JRS 61 (1971) 80—101; see also his comments on his earlier position in 
idem, Authority and the Sacred (Cambridge 1995) 57—78; and in Journal oj Early Christian 
Studies 6 (1998) 358-76. 

280 The standard book is L. Bieler, 0EIOI ANHP: Das Bild des ,Gottlichen Menschen' 
in Spdtantike und Fruhchristentum (Vienna 1935-36, repr. Darmstadt 1967), but it now 
seems very dated. 

281 H.-D. Betz, Lukian von Samosata und das Neue Testament: Religionsgeschichtliche und 
paranetische Parallelen. Ein Beitrag zum Corpus Hellenisticum Novi Testamenti (Berlin 1961) is 
a good account of Lucian's view of Christianity. 

282 Like all the 'divine men', Alexander of Abonouteichos has been the subject of 
a large bibliography, some of which may be listed here: A.D. Nock, Alexander of 
Abonuteichos, CQJ11 (1928) 160-62; M. Caster, Etudes sur Alexandre ou le faux prophete de 
Lucien (Paris 1938); L. Robert, A travers I'Asie Mineure: Poetes et prosateurs, monnaies grecques, 
voyages et geographic. BEFAR 239 (Paris 1980) esp. his lucid comments on Lucian pp. 
393—436; D. Clay, Lucian of Samosata: Four Philosophical Lives (Nigrinus, Demonax, 
Peregrinus, Alexander pseudomantis), ANRWII.36.5 (1992) 3406-3450; G. Sfameni Gas- 
parro, Alessandro di Abonutico, lo 'pseudo-profeta', owero, come costruirsi un'identita 



SYSTEMS OF BELIEF 117 

the mysteries naturally aimed to win a decent 'market-share' in this 
sector. The peace of mind granted by the miraculous activities of the 
oriental divinities of course predisposed others to join. In fact it was 
hard to choose between them in the holy shopping-malls sustained by 
the Empire's free market in religious goods (North 1992; 2003). It was 
the implicit aim of these cults to reduce the distress caused by this 
freedom of choice, though in the end the adaptability and ambiguity 
of mainstream Christianity won the day 283 

At the margin of institutionalised religion, however, magic proved a 
very effective means of social control. 284 Both its potential subversive - 
ness and the contempt felt for it by established intellectuals derive from 
this. The mysteries pushed their way into this market for personal fears 
too, since their claim to control of the universe applied equally to the 
local and contingent. Perhaps this is the reason for the links between 
Mithraism and magic, at least in the documents from Egypt, such 
as the so-called 'Mithras liturgy' (Betz 2003). On the other hand, it 
should be stressed that magic is an autonomous area, in the sense that, 
although it was of course shot through with elements borrowed from 
organised religion, including the mysteries, it aimed to create recipes 



religiosa, 1: II profeta, 'eroe' e 'uomo divino', SMSR 62 no. 20 (1996) [1998] 
D. Sabbatucci, 565-90; ~ II: L'oracolo e i misteri, in Bonnet and Motte 1999, 275-305. 
On Peregrinus note: R. Pack, The 'volatilization' of Peregrinus Proteus, AJP 67 (1946) 
334f.; G. Bagnani, Peregrinus Proteus and the Christians, Historia 4 (1955) 107—112 
(who strangely thinks he was an Ebionite Essene); S. Benko, Pagan Rome and the Early 
Christians (London 1985) 30—53 argues he was a marginal Christian. Apollonius of 
Tyana: the place to start is now C.P.Jones' excellent new Loeb ed. and transl. (Cam- 
bridge MA and London 2005); cf. G. Petzke, Die Traditionen iiber Apollonius von Tyana 
und das Neue Testament (Leyden 1970); E.L. Bowie, Apollonius of Tyana: Tradition and 
Reality, in ANRW II.16.2 (1978) 1652-99; M. Dzielska, Apollonius of Tyana in Legend 
and History (Rome 1986); J. -J. Flinterman, Power, Paideia and Pythagoreanism (Amsterdam 
1995); JA. Francis, Truthful Fiction: New Questions to Old Answers on Philostratus' 
Life of Apollonius, AJPh 119 (1998) 419-41. 

283 On the image of Christ against this background, see H. van der Loos, The Miracles 
of Jesus (Leyden 1969); A. Vogtle, The Miracles of Jesus against their contemporary 
Background, in H.J. Schultz (ed.), Jesus in his Time (Philadelphia 1971) 89-102 [= Eng. 
tr. of Die £eit Jesu (Stuttgart 1966)]; M. Smith, Jesus the Magician (New York 1978); 
B. Blackburn, The Theios Aner and the Markan Miracle Tradition (Tubingen 1991). I may 
here call attention to the debate in Lactantius, Inst. Die. 5.3.9 over the superiority of 
Christ or Apollonius, where he refers to a lost work by Hierocles Sosianus, the gov- 
ernor of Bithynia, who claimed that, if one had to adore a man, better Apollonius 
than Christ, since his miracles were finer and, on being arraigned before Domitian, he 
simply disappeared, whereas Christ was incapable of eluding crucifixion. 

284 R. MacMullen, Enemies oj the Roman Order: Treason, Unrest and Alienation in the Roman 
Empire (Cambridge, MA 1967). 



118 CHAPTER TWO 

'from outside' (formulae, phylacteries and amulets) that were intended 
to meet the same demands that the mysteries met by means of their 
own complex rituals. 285 

Not only did Isis offer Lucius the ass eternal life under her protec- 
tion but also a brilliant career here on earth as an advocate. Whatever 
Apuleius' true intentions may have been, I am interested in the way 
he ensures, despite the comedy, that his readers grasp the contempo- 
rary longing for the control of the real world that the Egyptian deity 
offers. She may also have recourse to magic to achieve her objectives, 
calming her adherents in the face of the hierarchy of powers, that is, 
her control of the range of means available to foresee and control the 
future. 286 That this was no trivial matter is clear from the eagerness of 
Christian writers in pointing out the inconsistencies of these beliefs and 
how the devil had inspired the related practices. If there had been no 
widespread interest among people in the Empire, such writers would 
never have expended such energy to prove that Christ's message was 
superior to any pagan claim and, above all, how he gave his followers 
a peace of mind that demonic superstitions could never afford. On the 
other hand, one just has to read the texts to realise how closely related 
all these various ideological productions were, that sought to assuage 
the fears aroused by the religious thinking of the time. The fears and 
the responses alike occupied the same cultural space. 287 

We unfortunately lack the evidence adequately to discuss one of the 
most universal aspects of religion, the regulation of sexual relations. 
This represents one of the basic concerns of all religious traditions, 



285 The relation between Mithraism, magic and the magi, esp. in the context of 
magical gems, is the subject of A. Mastrocinque, Studi sul Mitraismo (II mitraismo e la 
magia) (Rome 1998). He also delivered a series of lectures at the Ecole Pratique in 2007 
on the same topic. See also E. Sanzi, Mithras: A Deus Invictus among Persia, Stars, 
Oriental Cults and Magical Gems, Res Orientates 14 (2002) 209-29. 

286 In Egypt, Isis was closely connected with (temple) magical practice, for example 
in the historiola of Isis and Ra, used for healing snake-bites: in order to extract from 
ageing Ra his secret name, she makes a snake and smears its fangs with Ra's saliva. 
The snake bites the Sun-god, who thereupon implores Isis to help; which she gives, 
on condition that he reveal his secret name (P.Chester Beatty II recto 1.3/5; RTorino CG 
54052; cf. Mostra hide, 45 Cat. II. 2). Such historiolae were written on papyrus, dissolved 
in beer or wine, then drunk (see also Chap. 4.4.b.ii below). 

287 Cf. J.Z. Smith, Towards Interpreting Demonic Powers in Hellenistic and Roman 
Antiquity, AMRW 11.16. 1 (1978) 425-439. I think that it is this polyvalence of the 
mystery divinities in situations where traditional religion might have opted to combine 
gods of different kinds that lies behind Artemidorus' claim that the Nile deities have 
different functions in their physical doctrine, i.e. in their allegorical equivalents, and 
in myth (Oneir. 2.39 — perhaps also Cybele). 



SYSTEMS OF BELIEF 119 

since it is closely bound up with kinship-relations in society. Together 
they underwrite the ability of a society to reproduce itself in a culturally 
acceptable manner. The role ascribed to woman in sexual relations is 
crucial here, for example with regard to access to the marriage-market. 
The refinements of the rules of social conduct are extremely important, 
and it is here that Christianity had an enormously disruptive effect upon 
the Classical tradition. 288 The same cannot be said of the mysteries. One 
of the central themes of the cult of the Mater Magna was the issue of 
the incest-taboo, but it is bound up with the sacrifice of the part for the 
whole as an image of agricultural renewal. 289 The surviving evidence 
for the myth does not allow us to establish the acceptable rules with 
any clarity, though the taboo itself was clearly stated, as well as the 
penalty for its infringement. Nevertheless forgiveness was also possible, 
since, once the debt had been paid through self-castration, Attis, now 
pure, ended up at Cybele's side ([Hippolytus], Ref. 5.8.24). 

We have also seen how the myth of Osiris is imbued with sexual 
imagery and differentiation. An antithesis between the sexes is char- 
acteristic of Mithraism: no women are recorded as making votives, in 
sharp contrast to the other two mysteries, and, though images of female 
deities are not unknown, efforts seem to have been made to refuse 
the female principle, sometimes even by identifying natural women 



288 There were however tendencies within the pagan tradition itself on which this 
Christian strictness surely drew. Soranus for example claims: "Men who keep chaste 
are stronger and larger than others and keep healthier during their lives" {Gyn. 1.30.2); 
Galen says that Olympic athletes who are castrated are stronger (De semine 1 .8); Artemi- 
dorus records that an athlete dreamed that he cut off his testicles, bandaged his head 
and was crowned as victor. . .while he kept chaste he had a brilliant and distinguished 
athletic career; but when he began to have sexual relations again, he ended his career 
in obscurity (Oneir. 5.95); Quintilian suggests the advocate should refrain from sexual 
relations in order to acquire a deep masculine voice for the courts (Inst. or. 11.3.19). 
I am grateful to C. Martinez Maza for these texts. See also, more generally, Rous- 
selle 1983 and Brown 1988. U. Ranke-Heinemann makes the point succinctly: "It is 
untrue that Christianity introduced self-control and asceticism into a licentious and 
hedonistic heathen world. Sexual pessimism and hostility towards the pleasures of the 
flesh are a legacy from the ancient world which Christianity has preserved in a special 
measure to this day": Eunuchs for the Kingdom of Heaven.The Catholic Church and Sexuality 
(Harmondsworth 1990) 1. A more idealistic account can be found in F. Watson, Agape, 
Eros, Gender: Towards a Pauline Sexual Ethic (Cambridge 2000). More to my taste are the 
remarks of LP. Culianu, The Body Re-examined, in J.M. Law (ed.), Religious Reflections 
on the Human Body (Bloomington 1995) 2~ 4. C.J. Bleeker has drawn up a summary list 
of sexual themes that have been of interest to religions: Sexuality and Religion, in 
idem 1975, 208-24. 

289 Cf the account of the idea of sacrificing the part for the whole in order to obtain 
a greater benefit, though not with reference to agriculture, in Burkert 1996, 40—47. 



120 CHAPTER TWO 

with hyenas, an animal considered disgusting both in the Classical 
tradition and in early Christianities (e.g. Epistle of Barnabas ['Apostolic 
Fathers'] 10.7) because it symbolised the ambivalence of gender iden- 
tity and the difficulty of imposing the hegemony of the male. 290 The 
ideal cameraderie of those who shared the same gender was radically 
threatened by such disconcerting alterity Masculine insecurity could be 
sublimated through the domination of the other sex, as we can see in 
our own cultural system or can recognise in foreign systems that seem 
more transparent. 

One example here is the story of how the Dogon of south-eastern 
Mali acquired their masks, which were among the means by which 
the Andumbulu obtained power over others (Griaule 1963, 59-61). 291 A 
woman by chance saw the Andumbulu dancing in wooden masks rep- 
resenting an old man; through a trick she got hold of their magical 
objects, including the mask(s), and returned with them to her village. 
Everyone was terrified. She hid the objects in her store-room and got 
on with her tasks; but someone told her husband to take them away 
from her, for she had become a threat to men. Dressed in the mask and 
the other objects, he was able to thrash her. "The men, seeing that the 
objects belonging to the Andumbulu were an instrument of domination, 
decided to take them away from the women. 292 For up to that time, 



290 Gordon 1972, 98; 1980, 42-61. People have often tried to deny this, e.g. Mac- 
Mullen 1981, 203 n. 34; A. Blomart, Mithra: quoi de neuf en 1990?, JRA 9 (1996) 
435; J. David, The exclusion of women in the Mithraic Mysteries: ancient or modern?, 
Numen 47 (2000) 121—141 (not a serious contribution, in my view). The latest example, 
A. Griffith, Completing the Picture: Women and the Female Principle in the Mithraic 
Cult, Numen 53 (2006) 48-77, simply reinforces the point that women were not 
admitted into the cult of Mithras. Even if Gordon overstated the case regarding the 
elision of the female principle, the very occasional presence of female deities such as 
Venus or Victoria obviously, in the absence of dedications to Mithras by women, does 
not legitimate the conclusion that women were admitted to the cult. The Virunum 
album (AE 1994: 1334) is surely decisive here: out of 98 there is not a single woman. 
In this context, the contrast with the cults of Isis and Cybele, where such dedications 
are common, is very striking. Moreover the revisionists have themselves somehow to 
dispose of Pallas' claim about the hyena (ap. Porphyry, De abstin. 4.16.3), which they 
generally do with a nourish in the direction of textual corruption, thus rewriting the 
text to suit themselves. 

291 [The Andumbulu in Dogon lore were the first created people, who were like pyg- 
mies; they still live among the rocks in Mali but are invisible. Tr.] 

292 [Xhis quotation, which generalises the situation, is inserted into the narrative of 
the discovery of the mask by Griaule, apparently from another context. In a footnote, 
he refers to the Dogon creation myth, narrated on p. 46: in early times, the Dogon 
believed, women had had authority over men, which is why the soul after death goes 
first to one's mother's family. Tr.] 



SYSTEMS OF BELIEF 121 

the women in fact lorded it over the men, actually bullying them, and 
every day things got worse. If this state of affairs had gone on, men 
would have become the slaves of women." 293 

Investigation of the images of sex in the mysteries has to proceed 
indirectly, for example by looking at sensory perception, which would 
take me too far afield here, and anyway has already been studied. 294 
Moreover, the issue of sex goes beyond the myths, which are our 
immediate concern, and can be found not merely in the Mithraic 
rejection of women but also in the occasional accounts of sexual rela- 
tions, mentioned by (hostile) literary sources, considered improper for 
the adherents of the mysteries. 

For those aspiring to admission, the process of learning about the 
place of mankind in the world involved a period of preliminary prepara- 
tion during which part of the mystery was revealed. This is one aspect of 
a wider Hellenistic-Roman interest in wisdom or understanding (gnosis), 
which became a central theme in philosophico-religious speculation of 
the time. Access to the supreme being, to the One, could be attained 
by many routes, such as philosophy, divination, magic, mythology, ini- 
tiation into the mysteries, and so on. Each individual tried the route 
appropriate to his capacities and inclinations. 295 The mysteries were thus 
able to offer satisfying answers to the anxiety provoked by the central 
question, the meaning of life. 

I have argued that the mysteries provided a theodicy that legitimated 
the status quo by offering to those willing to observe certain ethical 
norms a vision of another world after death by way of compensation 
for this one. However that may be, the basic message of the myths 
for the individual, what gave the established order of things an air of 
immutability, capable of overcoming chaos, and believers the feeling 



29:5 The story suggested to me the idea of voluntary acceptance of female domination, 
a fiction rooted in male fantasies, which I have elsewhere (Alvar 1999) ironically called 
'the vice of dependency'. On the Dogon belief that the female sexual parts are dead, 
see G. Calame-Griaule, Ethnologie et langage: la parole chez les Dogon (Paris 1965) 296ff 

294 Cf. T. de la Vega, El aroma de los misterios, DHA 25.2 (1999) 41-54. She 
shows how the gradual integration of the mysteries into the cultural context of the 
Graeco-Roman world can be traced through the connotations of the scents associated 
with them in literature; cf. J. Alvar and T. de la Vega, La ambigiiedad cromatica en 
los misterios, in P. Ortega, MJ. Rodriguez and C.G. Wagner (eds.), Mujer, Ideologla y 
Poblaaon, ARYS 1 1 (Madrid 2000) 49-60. 

29:1 Cf. A.J. Festugiere, Cadre de la mystique hellenistique, in Aux sources de la Tradition 
Chretienne. Melanges Goguel (Neuchatel 1950) 74—85, offers a schematic but marvellously 
clear and richly nuanced account of the development of mysticism. 



122 CHAPTER TWO 

of being able to cope with their anxieties, was that everything is under 
divine control, and that it is in this that existence finds its meaning. 
Way beyond the issue of biological reproduction, the mysteries gave 
their adherents an irrational sense of things hanging together that took 
away the bitterness of the thought that we are merely creatures alive 
in the world. Small comfort no doubt in the face of the other central 
anxiety that I turn to next. 

3. The World Beyond 

Un Andoumboulou du nom de Golomine avait une 
fille, Mere, mariee avec un nomme Atamou. Un 
jour, le dieu Amma, qui peut prendre toutes formes 
et qui n'est pas connu des autres etres, amena une 
vache chez les Andoumboulou et la leur proposa. A 
Mere, qui en demandait le prix, Amina repondit: 
«Je la vends pour la Mort». Le femme, dans son 
ignorance, accepta le prix et conduisit la vache a 
sa maison. Atamou, son mari, mourut peu apres et, 
Mere, comprenant alors ce qu'etait le prix demande 
par Amma, se rendit aupres de lui pour lui rendre la 
vache. Amma refusa. La mort se repandit alors chez 
les Andoumboulou. 

Griaule 1963, 56 

Ai8oi) ... to xox> pioi) 7tepoc<; ixXvnov xe kou e\)86Ki|iov 
\iexa xr\q &ya0fj<; eXniSoq xr\q en\ xf\ nap' \>[iaq 



7iopeia/ 



Julian, Or. 5. 180c 



My main argument, reviving a major theme of the old grand narrative, 
has been that a soteriological promise was of fundamental importance 
in the oriental cults. I have briefly explained the problems involved in 
defining exactly what that salvation consisted in. But what we cannot 
reconstruct is the complex of beliefs regarding the world beyond this 
one. We have nothing comparable to the Christian controversies over 
the nature of eternal life. In our case, the dispute is limited to the 
question whether the mysteries really did posit a life beyond the grave 
closely calibrated with the fulfilment of the individual's religious and 



296 "Q ran ( xjaai. thg c l ose f m y lif e m ay be painless and glorious, in the good 
hope that it is to you, the gods, that I journey!" (tr. W.C. Wright). 



SYSTEMS OF BELIEF 123 

moral obligations. The debate concerns the meaning of the notion of 
salvation, ocoxripia or salus, used in Classical texts in relation to the 
gods of the mysteries. 

There has been an intense debate over whether ocoxripioc or salus 
relate to one's fate after death. Briefly stated, the issue is whether the 
offer of salvation was understood as happiness in the course of one's life 
after initiation, or as a promise of hope after biological death. Walter 
Burkert has flatly denied this: 

The Frazerian construct of a general 'Oriental' vegetation god who 
periodically dies has been discredited by more recent scholarship. There 
is no evidence for a resurrection of Attis; even Osiris remains with the 
dead . . . There is a dimension of death in all of the mystery initiations, but 
the concept of rebirth or resurrection of either gods or mystai is anything 
but explicit (1987, 75). 

This view seems to me quite mistaken. Let us take the evidence cult 
by cult. As regards Isis, the implications of the well-known account 
of Lucius' initiation into the mysteries of Isis, despite its deliberate 
reticence, have seemed to most scholars quite indubitable: 

Igitur audi, sed crede, quae vera sunt. Accessi confinium mortis et, calcato Proserpinae 
limine, per omnia vectus elementa remeavi; nocte media vidi solem candido coruscantem 
lumine; deos inferos et deos superos accessi coram et adoravi de proximo. 291 

Apuleius, Met 11.23 

So clear, indeed, that the discussion has revolved mainly around the 
cult of the Phrygian gods and Mithraism. 298 But it worth remaining 
with the Egyptian cults for a moment. The relation of these divinities 
to destiny was of course extremely complicated. Isis is herself destiny, 
in that she represents the cosmic order (through her association with 



297 "Therefore listen, but believe: these things are true. I came to the boundary of 
death and, having trodden on the threshold of Proserpina, I travelled through all the 
elements and returned. In the middle of the night I saw the sun flashing with bright 
light, I came face to face with the gods below and the gods above and paid reverence 
to hem from close at hand" (tr. J. A. Hanson). 

298 See esp. Brenk 1993, which usefully exposes some of the weaknesses of the nega- 
tive view. The continuities between pre-Hellenistic and Graeco-Roman Isis in relation to 
salvation have been traced by C.J. Bleeker, Isis as Saviour-Goddess, in S.G.F. Brandon 
(ed.), The Saviour God. Comparative Studies in the Concept of Salvation presented to E. 0. James 
(Manchester 1963) 1-16; cf M. Malaise, La survie dans les cultes isiaques, Acta Orientalia 
Belgica 3 (1983) 102-10. On the other hand, there is no evidence, positive or negative, 
before the time of Apuleius that Isiac salvation was thought of as other-worldly. Heyob 
1975, 60—64, however, argues that there are hints of a conception similar to that of 
the Metamorphoses in the funerary iconography. 



1 24 CHAPTER TWO 

the goddess of cosmic justice and order, Ma'at) and the regularity of 
natural processes. 299 Hence her identification with Tyche, Fortuna, which 
is one of the commonest of her iconographic types (Magris 1985, 508) 
(PL 4). Isis proclaims in hexameters at Andros that Seoncov 8' aeKovaocv 
<av>dyiccxv I avA/6a>, 'I untie the bonds of Necessity, reluctant though 
she be', a claim that is usually, and rightly taken to refer to the necessity 
of death. 300 In the 'Exchange of Moirai' text, Serapis claims that he 
can change men's Fates just by switching the goddesses' clothes: [xa<^\ 
Hoipocq yap eyro n£Tocrj(pid^co. 301 

Artemidorus recounts a particularly interesting story in this connec- 
tion (Oneir. 5.94). A man who was about to undergo an operation on his 
scrotum prayed to Serapis. The god appeared to him in a dream and 
told him that he should allow himself to be operated on without being 
afraid, for he would be cured. But instead he died. Artemidorus says 
that it was his destiny not to have to bear pain, just as if he had been 
cured, and adds that his end was something desirable, since Serapis 
is a god of the underworld (he lists him in his catalogue of such gods 
at 2.34). He does not make fun of the god's failure but looks for an 
alternative explanation to validate the prediction, despite the fact that 
it was wrong. Less straightforwardly, we could say that, whereas both 
the patient and Artemidorus wanted Serapis to intervene as a saviour 
(cf. Oneir. 2.39), the truly religious interpretation, and the reason no 
doubt why the story reached Artemidorus' ears in the first place, would 
be that, in place of a mere worry about locative, here-and-now salva- 
tion, the divine boon of Utopian salvation has been graciously granted. 
In the second half of the second century AD, both readings of the 
incident were possible. 302 



299 See for example a gilt-bronze statuette of XXII XXIII dynasty showing Ma'at 
squatting in front of Isis and Osiris: Museo Egitto, Turin Cat. 514; photo in Mostra 
hide 54 Cat. no. 11.15; cf. Griffiths 1970, 534. 

300 IG XII.5, 739 = Totti no. 2 = RICIS 202/1801 1.1 44f. with H.S. Versnel ap. 
Bianchi and Vermaseren 1982, 524. The earlier reference (1.96f.) to bonds being freed 
refers to the power of kings. 

301 PBerol. 10525 1.13 = Page, Select Papyri 3: 424-9 no. 96 = Totti no. 12 (IIP); see 
n. 10 above. 

302 For the terms, see n. 7 above. Vinagre 1994 has argued that Artemidorus seems 
to treat curative dreams sent by Serapis and Asclepius differently: the former, as a new 
divinity, was able to make use of new forms of publicity made possible by the spread 
of books, and therefore looms large in the oneirocritic literature. 



SYSTEMS OF BELIEF 125 

However, our evidence regarding the Egyptian cults is exception- 
ally rich. 303 If we had the same quantity of documentation about the 
others there would surely have been no need to spill so much ink over 
the issue of salvation. In the case of the Phrygian cults, Giulia Sfa- 
meni Gasparro argues that, although no source explicitly tells us that 
the initiates were given guarantees of immortality, the theme was not 
entirely absent; however she claims that it certainly was not constitutive 
of this type of religion. 304 More restrictive still is the position of Ileana 
Chirassi Colombo, who thinks that the salvation of the soul was not 
a feature of the mysteries, and that the notion of salus was far more 
concrete and contingent than is often assumed. It refers primarily to 
an attitude to this world, based on sacrifice, performed in or for the 
god. It also involves coming to accept the 'impossible' world as the 
goal of escape from this one, as suggested by the example of Attis and 
the ritualism controlled by the clergy (Chirassi Colombo 1982, 326). 
In view of the evidence I shall be considering, this seems to me an 
extremely reductive account. 

As for Mithras, it is Robert Turcan who has been most sceptical of the 
traditional idea, though in his later work he seems to have moderated his 
criticism (Turcan 1992a). The question revolves around the interpreta- 
tion of 'the ascension of the soul', the heavenly journey mentioned by 
Celsus and Porphyry 305 To put the matter simply, Turcan rejects these 
statements on the grounds that they are not actually from informants 
who were initiates into the cult but simply later claims by neo-Platonists. 
They are therefore not primary evidence, merely distractions. According 
to him, individual salvation was not a feature of Mithraism: the bull's 
death constitutes a once-and-for-all 'bio-cosmic' salvation. Others, such 
as Roger Beck (1988, 77-79; 2006, 102-12), reject this and believe that 
Celsus and Porphyry do tell us about actual Mithraism, but that their 



303 It is these cults too that have most evidence for benefits in this world and for 
divine aid. We may for example cite the miraculous triumph of Apollonius in the case 
brought against him for having built a temple of Serapis (Serapeum A) on Delos in 
the late third cent. BC (IG XI.4, 1299 = Longo 1969, 106-16 no. 63 = Totti no. 1 1 
= RICIS 202/0101, with Engelmann 1975. 

304 Sfameni Gasparro 1985, 123; in eadem 1982, she comes even closer to the posi- 
tion of Chirassi Colombo. 

305 Celsus, ap. Origen, Contra Cels. 6.22 — Sanzi 2003, 416: Mithras no. 7; Porphyry, 
De antra 24 - Sanzi 2003, 422: Mithras no. 8.7. 



126 CHAPTER TWO 

information relates to the process of initiation in the mithraeum, and 
that the world beyond is not the primary issue here. 306 

I incline to think that an account of the soul's journey was given 
during initiation, as the sole means of presenting vividly a doctrine 
whose truth the initiand would later have an opportunity to experience 
first-hand. The ritual as performed in the mithraeum did contain some 
kind of soul-journey towards perfection that, according to Porphyry {De 
antro 24), took place under the guidance of Cautes. We all know, how- 
ever, that initiation is an imitation of death, so it is not surprising that 
Mithraism should have made the point in these terms. I am attracted 
by the argument that there were two imaginary worlds beyond this one, 
one concerned with initiation and the other with the world after death. 
Although we have no specific knowledge of this latter, I am convinced 
that the Mithraists believed in it (cf. Merkelbach 1984, 244). My belief 
is that Mithraic initiation liberated the soul, and that this liberation had 
to be repeated at the point of biological death. The purificatory rituals 
that accompanied initiation into the different grades were intended to 
keep the soul unsullied for this final journey. What we cannot tell is 
whether this was a belief known only to neo-Platonist Mithraists with 
a philosophical background, or whether it was a widespread claim 
independent of the educational and intellectual background of the 
individual adherent (cf. Gordon 1988, 45). 

On the other hand, the problem of Julian's allusion, at the very 
end of his Banquet of the Caesars, to Mithras' role as psychopomp, that 
is, his role in guiding souls to their final destiny (336c), cannot be so 
easily resolved. 307 Turcan, still following his argument that all this is 
neo-Platonist or -Pythagorean interpretation, argues that Julian's Father 
Mithras is not our Mithras but the syncretistic Helios worshipped by 
the emperor, and hence that what he says here is irrelevant for our 
understanding of true Mithraism. 308 We have to allow that beliefs and 



306 Beck 1996, 132 accepts salvation in the world beyond, but contrasts it with 
Christian salvation, without giving details. 

307 Kai fiviKa av ev0ev8e &7uevou 8er|, |xexa xr\c, ayaGfji; zknihoc,, r|ye|j,6va Geov 
£t>|xevfj KaGtotai; aeceuiS, cf. B. Lincoln, Mithras as Sun and Saviour, in Bianchi and 
Vermaseren 1982, 505-23 at 506f. 

308 Turcan 1981a, 113 = 2000, 112, but see ibid, 145-52. R. Smith adopts a more 
moderate position in Julian's Gods: Religion and Philosophy in the Thought and Action of Julian 
the Apostate (London and New York 1995) 137, where he affirms thatjulian was initiated 
into the Phrygian mysteries, but remains sceptical about Mithraism. He never however 
expresses doubts about the Mithraic character of some of the material. 



SYSTEMS OF BELIEF 127 

cults alter under pressure of the historical reality in which they exist, 
it is therefore perfectly in order to think that the Mithraic conception 
of the world beyond may have changed over time. On the other hand, 
from the late second century AD Helios/Sol acquired a variety of fea- 
tures taken over from sun-gods with specific epithets in the same way 
that he assimilated aspects of other deities with solar connotations. The 
root of the problem of course is how we are to distinguish between 
the different solar cults. Everyone here draws his own more or less 
'optimistic' conclusion (Liebeschuetz 1999). Julian's psychopomp may, 
I think, carry traces of Plutarch's mediating Mithras {De hide 46, 369e), 
whom no one now takes very seriously because he does not seem to fit 
into what we know of Roman Mithraism. He was interpreted in the 
Cumontian tradition as a trace of the god who acted as an intermedi- 
ary between the opposing poles of Iranian dualism; but it may be that 
Plutarch believed he had the same mediating function that he has in 
Roman Mithraism, namely as a demiurgic instrument of creation and 
salvation, and consequently, as Julian claims, a psychopomp. Plutarch 
locates the problem in Zoroastrianism, because of its prestige, but may 
in fact be referring to Mithras as a mediator of the initiate's soul, who 
saves it from the darkness of Hades and guides it to an eternity filled 
with light. 

The Mithraic heaven was divided into two halves, a sort of map 
inscribed with coded meanings, a visible mediation between this world 
of mutability and death and the other of purity and immortality, where 
the heavenly bodies became metaphors of salvation, and astronomy 
an indispensable key to the ciphers. 309 As we have seen, according to 
Porphyry Mithras' seat was at the equinoxes, which is an intermediate 
position: the equinoxes are not merely the points at which the ecliptic 
crosses the celestial equator, but the moment when day and night are 
exactly equal in length all over the world. Mithras thus occupied and 
controlled the mid-points between the summer and winter solstices, the 
axes of the year. Porphyry continues in the same passage to say that 
Mithras therefore has north to his left and south to his right. North, 
which is cold, is the place of Cautopates and coming-into-being [genesis), 
south, which is warm, that of Cautes and going out of being (apoge- 
nesis). One of the Mithraic interpretations of the torchbearers seems 



309 Cf. Gordon 1975, 226; 1988, 49; Beck 1994, 30; somewhat differently, 2006, 
passim. 



128 CHAPTER TWO 

thus to have been that they marked the process of souls coming into 
being (Cautopates' lowered torch) and going out of being (Cautes with 
his torch raised towards heaven pointing to the gate through which 
the souls pass on their stellar journey to eternity made possible by the 
victory of tauroctonous Mithras). 310 

I do not think it surprising that Mithras should have possessed aspects 
related to the beliefs of the time regarding the world after death, rooted 
as they were both in philosophical reflection and in current anxieties. 
Right from the beginning, the god was linked to justice. As guarantor 
of the order of the cosmos, we should think of him as presiding over 
destiny. If we take into account his 'heroic' aspect that was so much to 
the fore in the Hellenistic-Roman period, 311 we can think of him as a 
veritable model on whose basis men could build and attain their own 
salvation, in that the hold of fate over the individual, which we must 
assume to have been a theme in Mithraism too, could be loosened 
through following the cult's ethical demands. At the same time, however, 
Mithraism did have, as one of its many paradoxical features, a special 
conception of pre-destination in the form of the figure of Time (also in 
the form of Aion), which played a considerable, though to us completely 
obscure, role in its organising discourse (PI. 8). Individual salvation may 
have involved a form of predestination that could apparently be over- 
come thanks to the initiatory process. It was this latter that overcame 
belief in the absolute control of Fate (Magris 1985, 508ff). 

On the other hand, Porphyry does claim, citing Euboulus as his 
authority, that the highest class of Persian magi abstained from eating 
meat on the grounds that one of their most important doctrines was 
metempsychosis 'which also seems to have been the case in the mysteries 
of Mithras'. 312 He goes on to link the association of Mithraic grades 
to animals to this idea. If we can believe this, salvation in Mithraism 
would have been correlated with the grade of the soul's re-incarnation. 
There is, however, no supporting evidence, which explains the extreme 
scepticism of modern scholars about this claim. 313 It is perfectly pos- 



310 Cf. the useful chart given by Beck 2006, 210 fig. 13 (and pp. 209-14 as a 
whole). 

311 Gordon 1988, 60-64; Zwirn 1989. 

312 De abstin. 4.16.2 = Sanzi 2003, 418: Mithras 8.2. 

313 Particularly, once again, following Cumont, Turcan 1975, 30-33; 1981a, 112 = 
2000, 111. By contrast, Merkelbach 1984, 238-42 takes it as the most natural thing 
in the world. 



SYSTEMS OF BELIEF 129 

sible that Porphyry, or Euboulus, is here just attributing a neo-platonist 
doctrine to the adherents of Mithras, assumed to be committed to the 
same beliefs as the famous Persian magi, as part of a wider strategy of 
finding Platonist themes in barbarian wisdom. 314 On the other hand, 
although modern experts on Iranian religion have generally ridiculed 
the idea, Shaul Shaked has recently suggested, in a radical rethinking 
of the traditional picture of pre- and Sasanian Zoroastrianism, that 
there may indeed have been mobeds who entertained ideas of this kind, 
of whom Greeks under the Empire could have heard more or less 
reliable reports. 315 I do not think it would be surprising for a cult that 
believed in the descent of souls from the stars and their return also 
to claim that the next time around the very same souls could return 
to this world to be re-incarnated once more, repeating the experience 
of coming into being, as Porphyry says, whether or not under neo- 
Pythagorean influence. For the return or the eternal salvation could 
have been conditional each time on individual conduct, as we shall see 
in the following chapter. 

We are here, of course, faced with a problem we have come across 
several times already: how far we can believe a single item of evidence? 
Naturally people have different views on this sort of issue, but it seems 
to me that in any individual case the information may be true, at least 
in some sense; we have, though, as here, to understand it aright. I think 
it perfectly possible that, as Euboulus says, some Mithraists believed in 
re-incarnation as a variant of a wider belief, shared by all adherents of 
the cult, in the continued existence of the soul after death. In the present 
context, what is important is that by the third century AD Mithraism 
accepted the existence of the individual soul, with its own continued 
existence after death, influenced by one's personal conduct in life, and 
whose fate could be decisively affected by adherence to the cult. This is 
how I take Tertullian's reference to the imago resurrectionis in Mithraism, 



311 Bidez and Cumont 1938, 26 and 28; H. Dorrie, Kontroversen um die Seel- 
enwanderung im kaiserzeitlichen Platonismus, Hermes 85 (1957) 414—435; W. Deuse, 
Untersuchungen zur mittelplatonischen und neuplatonischen Seelenlehre (Wiesbaden 1983) 129-166; 
A. Smith, Did Porphyry reject the transmigration of human souls into animals?, RhM 
127 (1984) 276-284; cf. too the remarks of M. Patillon and P. Segonds in the Bude 
edition of De abstin. ad loc. 

51:1 S. Shaked, Dualism in Transformation (London 1994); idem, Popular religion in 
Sasanian Zorastrianism, Jerusalem Studies in Arabic and Islam 21 (1997) 103—17. 



130 CHAPTER TWO 

which of course is denied by all those who on principle refuse to believe 
the information provided by this Christian apologist. 316 

In view of all this, I would argue that the issue of salvation should 
not be viewed in the rather black-and-white terms that it often is. For, 
from the point of view of the mysteries, initiation and death formed 
a single process or trajectory, so that it is perfectly legitimate to think 
that salvation could have both locative and Utopian aspects. This may 
be the point to refer again to the passage, probably from Plutarch's 
De anima, cited in the previous section (2.2), and which we shall have 
to refer to yet again in relation to the ritual itself. Of course we do 
not know which mysteries Plutarch is here referring to, but the initiate 
clearly believes that, when he reaches the point of death, the experience 
is going to be similar to what he saw in the initiatory trance. 317 The 
'ascent of the soul' was thus — as Turcan rightly argues — experienced 
proleptically in the mystical ecstasy of initiation and then again in the 
actual dark journey towards death that replicated the ritual once and 
for good. 318 

Given that general context, the emergence of a grade-structure 
(whatever its form) and the attempt (at least in Rome and Ostia) to link 
it with the planetary system seem to suggest that at least in some places 
there emerged a Mithraic journey in which the initiates entered the 
gates of heaven, once they had been purified from this earthly life. 319 



316 Tertullian, De praescr. 40.4 = Sanzi 2003, 414: Mithras no. 5.2. This work, like 
De baptismo, was written c. AD 200, whereas the Contra Marcionem and De corona belong 
to his Montanist phase, i.e. after 207; cf. Beskow 1994, arguing that Tertullian wrongly 
interpreted the Mithraic cave and the rock-birth as a parody of the birth and resur- 
rection of Christ and not as authentic Mithraic rituals. 

317 Y. Verniere, Initiation et eschatologie chez Plutarque, in Ries 1986b, 335—352 
argues that the vision one has as one dies was somehow reproduced in the initiatory 
trance. In the case of the passage of Plutarch, it is highly likely that he is talking of the 
Eleusinian mysteries and not of the cult of Isis, as Dunand 1973, 3: 250ff proposed; 
cf. F. Graf, Eleusis und die orphische Dichtung Athens in vorhellenistischer /jit (Berlin 1974) 
132-38; Burkert 1987, 91-93. 

318 I may here refer to a couple of the studies publishd by the Donner Institute for 
Research in Religious and Cultural History: S. Hartman and CM. Edsman (eds.), 
Mysticism (Stockholm 1970), and N.G. Holm (ed.), Religious Ecstasy (Stockholm 1982), 
both of which have important introductions by the editors. Cf. too W. Wainwright, 
Mysticism: A Study of its Nature, Cognitive Value, and Moral Implications (Madison 1981); 
M. Laski, Ecstasy: A Study of some Secular and Religious Experiences (Bloomington 1967). 

319 The image at the entrance to the Mitreo delle Sette Porte at Ostia is quite sug- 
gestive in this regard: it consists of an arcade of six arches with a taller seventh in the 
centre (Becatti 1954, 96f with pi. XIX = V 287). Such an image can be linked with 
some Campanian plaques, theatre-scenes and of course the arcade-sarcophagi, where 
one sometimes finds that the central niche is in one way or another especially empha- 



SYSTEMS OF BELIEF 131 

The interference between Mithraism and gnosis /Platonism makes it 
hard to understand quite what the situation was but at the same time 
there is an important gain from setting the cult in the wider context 
of contemporary philosophical speculation. 320 

I might add that belief in the world beyond does not depend so much, 
in the real world, on personal religious inclinations, that is, on ideas 
theoretically available in the mystery context, but on what was taken 
as a fact in the general system of beliefs current at the time when the 
mysteries developed (cf. Engster 2002). 

In such a case, it is the wider belief, in this context, the cultural 
superstructure of the Graeco-Roman social-formation, that dominates 
the particular, here the mysteries. In a world that frantically looked for 
ways of resolving the conflicts produced by human ignorance of what 
happens after death, it would be incomprehensible if the mysteries, 
and Mithraism in particular, believed in nothing after this biological 
existence. Even if it had been possible metaphysically their adherents 
would have found it difficult to resolve the anxieties that were inevitable 
given that the wider culture entertained such troubling speculations 
about the world beyond. 321 



sised, e.g. M. Wegner, Die Musensarkophage. ASR 5.3 (Berlin 1966) nos. 6 (Aphrodisias); 
20 (Brussa); cf. 206 (Museo Borghese); the motif is esp. common in Asia Minor sar- 
cophagi, but cf. also P. Kranz, Zu den Anfangen der stadtromischen Saulensarkophage, 
MDAI(R) 84 (1977) 349-80. 

320 Tjircan dismisses all the information deriving from Porphyry's sources as con- 
taminated by neo-Platonism or -Pythagoreanism and thus of no use in telling us about 
genuine Mithraism (1975, 23—43). This view derives from his prior commitment to 
Cumont's theory of the Mazdean origin of the Roman mysteries. As Turcan himself 
admits, however, it is extremely difficult to discriminate here between the genuine 
and suppositious, given that the cult must have developed in the western part of the 
Empire in the second half of the first century AD. I think we should have no truck 
with ideas like 'pure' and 'genuine' in relation to cultural phenomena. R. Wagner, 
The Invention of Culture (Chicago 1980) is rightly critical of claims about 'purity' and 
'authenticity' in this area; following on from that, I see the problem rather in terms 
of the invention of tradition: cf. E. Hobsbawm and T. Ranger (eds.), The Invention of 
Tradition (Cambridge 1983). 

321 One remembers in this connection the famous passage of Parmenides (28 A46 
DK = Theophrastus, De sens. 1.3) where he describes the dead body as capable of 
experiencing light, cold and silence. Its extensive later echoes have been studied by 
R. Verdiere, Le concept de la sensibilite apres la mort chez les anciens, Latomus 50 
(1991) 56-63, who sets the idea of post mortem expectations into the mysteries' cultural 
context; cf. Burkert 1987, 23ff Some further reflections, not quite apposite here, but 
which aid in one's general thinking on these matters, may be found in H. Mathieu, 
Resurrection et immortalisation, in F. Jouan (ed.), Mort etfecondite dans les mythologies. Actes 



132 CHAPTER TWO 

Given this general context, it would not be surprising to find that, 
as Origen claims, some mystery systems accepted a belief in eternal 
punishment after death. 322 We need to keep in mind both the historical 
context, which created one complex of beliefs, and the historiographi- 
cal context, which created another within the frame constituted by the 
first. At any rate, I think we can properly claim that the institutions that 
claimed to speak for the gods, that is, constituted beliefs and the imposed 
social order taken together, taught that it was possible to obtain a just 
reward for patiently enduring the world's slings and arrows, namely a 
marvellous experience in the world beyond. Unfortunately, this desir- 
able condition could only be reached by dying. 

The offer of eternal happiness for initiates of the mysteries emerges 
from the myths. Frazer thought of the evocative term 'dying and rising 
gods', which has become the stereotyped definition. Critics however 
have pointed out that this is not so much an accurate description of 
them as a handy way of summarising their myths. I therefore prefer 
to call them 'gods of death and resurrection'. We can perhaps draw a 
very schematic contrast between the Homeric Hades and the myster- 
ies in this context. In Homer, the sole form of post-mortem existence 
is as a shade regretting no longer being alive, or thinking about what 
happened back there in the real world. 323 The mysteries on the other 
hand seem to have offered an agreeable eternity where the initiate 
became part of the god's company and shared his table. 324 Platonist 
and platonising speculations in the later Graeco-Roman world gave rise 
to analogous dreams (Merkelbach 1984, 228— 44). The mysteries were 
thus fully in keeping with the collective imaginaire of their age, which 
offered themes that were taken up both by the Olympian tradition and 
by the new movements generated by the process of historical change. 



du Collogue de Poitiers (Paris 1986) 39—49. Note too the quite different conclusions, in the 
context of the Endymion sarcophagi, of M. Koortbojian, Myth, Memory and Meaning on 
Roman Sarcophagi (Berkeley and Los Angeles 1995) 85-99. 

322 Contra Celsum 8.48; cf. E Cumont, Lux Perpetua (Paris 1949) 219. Lane Fox 1988, 
96 rejects the passage; in my view, though it may be contaminated with Christian ideas, 
it does give us a glimpse of the reality of the ideological shift that was taking place. 

323 Cf. J. Griffin, Homer on Life and Death (Oxford 1980) 90-95, 160-63; J.N. Bremmer, 
The Early Greek Concept of the Soul (Princeton 1983); Garland 1985, If., 68. 

324 Garland 1985, 68-72 suggests however, on the basis of grave-goods and Musaeus 
ap. Plato, Rep. 363cd = Kern, OF frg. vet. 4, where the just are rewarded in Hades 
by being able to indulge in continuous symposia, that it was widely believed in the 
Classical period that one might feast, drink, play draughts, perhaps even have sex in 
the Underworld. 



SYSTEMS OF BELIEF 133 

Triumph over destiny and admission to eternal felicity were however 
only possible through divine suffering. That is why the mysteries needed 
divinities who had had some experience of something like the human 
condition, had themselves lived historically so that they could function 
as models. Their adherents might suffer pain and torment, but with the 
god's aid they could overcome them. Their individual successes were 
partial victories over destiny (and potentially over the established order) 
and made them worthy to join the eternal company of the gods, the 
true architects of absolute victory. 

Some may doubt that initiates looked forward to feasting with their 
gods after death, but another passage from Apuleius is quite explicit 
on the point: 

Nam et inferum claustra et salutis tutelam in deae manu posita, ipsamque traditionem 
ad instar voluntariae mortis et precariae salutis celebrari, quippe cum transactis vitae 
temporibus iam in ipso finitae lucis limine constitutos, quis tamen tuto possint magna 
religionis committi silentia, numen deae soleat elicere et sua providentia quodam modo 
renatos ad novae reponere rursus salutis curricula. 32 '' 

Metam. 11.21.6f. 

It is of course not good method to project onto the other gods what 
Apuleius here says about Isis. Indeed his claims here may not even 
have applied at other periods even to her. On the other hand, there is 
epigraphic evidence that suggest that I am right about salvation in the 
mysteries in general. Some parents request the 'fresh water' (to \|/\)%p6v 
vScop) of Osiris for their dead children, which surely expresses the hope 
that they shall be immortal: the water is the sacred Nile-water fetched by 
Osiris. 326 Another text from Rome hopes that the dead person will rest 



325 "Yor (said the priest Mithras) both the gates of death and the guardianship of 
life were in the goddess' hands, and the act of initiation was performed in the manner 
of voluntary death and salvation obtained by favour. In fact, those who had finished 
their life's span and were already standing on the threshold of light's end, if only they 
could safely be trusted with the great unspoken mysteries of the cult, were frequently 
drawn forth by the goddess' power and in a manner reborn through her providence 
and set once more on the course of renewed life", tr. J.A. Hanson. 

326 T/he m0 st interesting case is Lehmann and Holum 2000, 139f no. 158 = RICIS 
403/0401 (Caesarea Maritima, where numerous Isiac funerary images were found in 
the 1990s): a father addresses two of his children who died more or less simultaneously, 
and expresses the wish that Osiris grant them both his refreshing water. The text ends 
with an interesting fusion between the traditional and the new: rfj fj|xTv eXcxippa tcai 
ra Kara 5ot><; aya0d {sic), 'may the earth lie lightly upon you and may the lot (that 
Osiris) grants you in the Underworld be good'. See also IGUR 836 = RICIS 501/0164 
(where the child has been taken 'not by Charon but by Chaos'); 432 — 501/0178; CII 
VI 20616 - RICIS 501/0198; cf. IG XIV 1842 (all Rome); SEG IX (1944-45) 829 - 



134 CHAPTER TWO 

in peace with Osiris (e , 6\|rij%i |a,exa xov 'OoeipiSoc;). 32 ' That the salvation 
offered Lucius by Isis is not just a promise that he will be fortunate in 
this life once he is initiated — provided he fulfils all the requirements — but 
extends to the world beyond seems clear from the opening phrase: Nam 
et inferum claustra et salutis tutelam in deae manu posita.™ 

There is unfortunately no evidence for the Phrygian cults or Mithras 
comparable to Book 1 1 of Apuleius' Metamorphoses, by means of which 
the doubts of scholars may be quieted. If they cannot even agree 
about the post-mortem promises of the Egyptian cults, it is only to 
be expected that disagreements over Mithraism should be fiercer still. 
As we have already seen (p. 125), the primary literary evidence for 
soul-journeys in the cult, albeit rejected by Turcan, is the passage of 
Porphyry, De antro 24 that I have already cited in relation to the seat 
of Mithras on the equinoxes, and Origen's long citation from the oth- 
erwise unknown philosopher Celsus, to which I have already alluded 
several times, concerning the opposition between the fixed stars and 
the planetary system, represented as a 'ladder' rather than a series of 
concentric circles. 329 The ladder represents, inter alia 'the passage of 
the soul between them', kou xfiq 8i' ouucov xr\q V^X^i^ §i£^6Sot> (cf. 
Beck 2006, 83f), i.e. between earth and the fixed stars. 

Apart from these texts, which one would have thought were explicit 
enough, much of the discussion used to revolve around what was taken 
to be the reading of Line 14 at Santa Prisca in Rome when the dipinti 
were first published by Vermaseren in 1965: et nos servasti eternali sanguine 



RICIS 703/01 1 1 (Carthage) with D. Delia, The Refreshing Water of Osiris, Journal oj 
the American Research Center in Egypt 29 (1992) 181-90; R. Rubio, Los isiacos y su mundo 
funerario, in Alvar, Blanquez and Wagner 1994, 162. At any rate down to the late 
Ptolemaic period, the (privileged) dead in Egypt were believed to be prepared by Isis, 
Nephthys and Anubis, for assimilation to Osiris, King of the Dead. 

327 IG XIV 2098 = IGUR 1042 = RICIS 501/0196 (lost). However ewraeiv is the 
standard farewell in all these texts, and does not in itself legitimate the inference that 
the person is to enjoy 'eternal life' in a strong sense. The notion of eternal home/ 
dwelling occurs twice in IGUR 836 = RICIS 501/0164 (see previous n.), but this trope 
simply means the person is dead and will never be seen again here in the world of 
the living. 

328 So rightly Griffiths 1982, 202. However in his commented edition of the Meta- 
morphoses, he translates: "The gates of hell and the guarantee of life were alike in the 
hands of the goddess...", which gives a quite false impression of both meaning and 
allusions (1975, 95). See also the discussion of Mimbu Kilol 1994 in support of a 
'documentary' reading of Bk. 1 1 against the ironic Winkler-Reardon-Hagg reading 
(cf. n. 278 above). 

329 Contra Cels. 6.22 - Sanzi 2003, 416: Mithras 7. 



SYSTEMS OF BELIEF 135 

fuso, 'You have saved us by shedding the eternal blood'. Vermaseren 
and Hans-Dieter Betz were of course struck by the analogies in the 
Christian liturgy; Turcan on the other hand saw only a once-for-all 
salvation in this world: the death of the bull was a unique event that 
liberated the demiurgic, salvific power of Mithras here on earth, which 
could not be repeated. 330 However Silvio Panciera showed a quarter- 
century ago, after the cleaning of the late 1970s, that the reading was 
the product of Vermaseren's wishful thinking, and bore little relation 
to anything that can ever have been visible on the wall. The reading 
eternali is quite impossible; every other word in the first half, including 
servasti, is very problematic; and only the phrase sanguine fuso at the end 
is reasonably certain. 331 That Mithras achieved something by killing 
the bull we knew anyway. The entire discussion can thus now only be 
viewed in the Musee de Idees Mortes. 

Since then, however, another fragment of possibly relevant evidence 
has come to light, the prefatory note to an important inscription found 
deliberately concealed in a Roman cellar in Virunum, Noricum. It refers 
to an occasion, 26 June AD 184, on which the Mithraists organised here 
held a ceremony to commemorate the death of some of their fellows, 
among them a Father, a member of the highest grade: qui templum vii 
conlapsum impendio suo restituerunt / 1 (added in smaller letters:) et mortalitatis 
causa conven(erunt) Marullo et Aeliano cos. VI Kal. Iulias. 7 ' 32 Due to circum- 
stances unknown to us, this ceremony coincided with the re-dedication 
of the mithraeum, which had been restored partly at the expense of 
the most prominent member, partly by contributions from the individu- 
als named. This text is in fact the sole documentary reference from 
within the cult to death: there are no clearly Mithraic tomb-stones 
(just funeraries for Mithraic priests set up by their wives), of course no 
sarcophagi, and, by contrast with the cases both of the Egyptian and 



330 Vermaseren and van Essen 1965, 217; Betz 1968, 77; Merkelbach 1984, 145£; 
199. This version is still cited even by those who ought to know better, e.g. Turcan 
2000, 109f. (omitting eternali, which anyway did not scan); Sanzi 2003, 439: Mithras 
no. 26.1, 1.14. 

331 Panciera 1979, 103—5 note **, with the drawing on the pull-out immediately 
before p. 127. 

332 Piccottini 1994 - AE 1994: 1334, cf. R.L. Gordon, Two Mithraic albums from 
Virunum, Noricum, JRA 9 (1996) 424—26. The death of the senior Father is indicated 
by the usual sigle, a Greek in the margin beside his name. Roger Beck has argued 
that the date was shifted deliberately so as not to coincide with the solstice, on his 
view the time when souls came into the world through Cancer (1998b; cf. 2006, 84; 
209-12). See also Chap. 4.5.b below. 



136 CHAPTER TWO 

Phrygian cults, no specifically Mithraic funerary imagery 333 But it does 
suggest that the deaths of members may have been commemorated on 
a regular basis, thus providing an opportunity to reinforce the claim 
about Cautes, heat, and apogenesis, going out of being 334 

With regard to the Phrygian mysteries, there is more consensus 
among scholars, despite the fact that there is no irrefrangible evidence 
for eternal salvation here either. 333 Since the days of Bidez and Cumont, 
a passage of Firmicus Maternus has often been cited in this connec- 
tion. It describes a ritual in which the statue of a god was buried at 
night, the adherents sung or wailed rhythmic lamentations, and, after 
their throats had been anointed with oil, the priest slowly spoke the 
following phrase in a low voice: 



©appeite pcbaxai xox> 0ecn3 aeacoaiievou 

Firmicus Maternus, De errore 22.1 



eotai yap fuiw rcovcov acotripia . . , 336 



This salvation can only be the triumph over Destiny and the promise 
of a life in the hereafter markedly different from that assumed by the 
dominant Graeco-Roman religious system (granted that it too changed 
over time). 337 The phrase seems clearly to be a formulation from the 
mysteries, but cannot be linked to any particular cult, though it is tempt- 



333 The absence cannot be connected with the lack of a dying god: the main funer- 
ary images in the cult of Isis are those of women wearing the 'knot of Isis': Walters 
1 988; J. Eingartner, Isis und ihre Dienerinnen in der Kunst der romischen Kaiserzeit. Mnemosyne 
Suppl. 115 (Leyden 1991). 

334 In ordinary professional and collegial associations, individual donors commemo- 
rated their birthdays by means of a feast supported by a foundation, not the day of 
their death or funeral. There is no epigraphic trace of such activity by Mithraists. 

335 There is a quite widespread desire to distinguish Mithraism from the other 
two cults, which, for unknown reasons, appear to bear a greater resemblance to one 
another. 

336 "Take heart, initiates of the god that has been saved, for there shall be salvation 

for us from our sufferings " = Bidez and Cumont 1938, 2: 285 no. 3. J. Podemann 

Sorensen, Attis or Osiris? Firmicus Maternus, De errore 22, in idem 1989, 73—86, argues 
that the symbolon refers to an Osirian ritual, because there are parallels in the Egyptian 
cults for almost all the points made by Firmicus. His conclusion may be doubted, since, 
as he himself admits, despite our relatively good information about the Egyptian cults, 
there is no other evidence for this belief being current there. 

337 Cf. still the epigraphic collections by R. Lattimore, Themes in Greek and Latin 
Epitaphs. Urbana Studies in Language and Literature 28.1—2 (Urbana 1942); W. Peek, 
Griechische Grabgedichte. Schriften und Quellen der antiken Welt 7 (Berlin I960); A.-M. 
Verilhac, Paides aoroi. Poesie funeraire (Athens 1978—82). 



SYSTEMS OF BELIEF 137 

ing to think of a day corresponding to the Hilaria at Rome. 338 That 
is why Turcan does not make use of it, even though he does argue, 
on the basis of certain images, and texts by Julian and Sallustius, that 
Cybele and Attis promised the triumph of souls over death, that is, a 
return to the divine. 339 Of course no one doubts that there were marked 
interferences between the mysteries and Christianity during the fourth 
century AD; but these ideas are perfectly in keeping with the beliefs of 
the mysteries at this period of close contact. After all, we find Julian 
himself asserting that the souls of the followers of Attis, so long as 
they keep his commandments, can leave the world of becoming and 
limitlessness, and return to the gods: xi Se iXapcoxepov yevouo ocv \|n>xfjc; 
OOTeipiocv |iev kocI yevemv kou tov ev oanfj kA,v>Scdvcx Sicx(pi)yo'6o"r|<;, in\ 
Se xovc, Beoi)!; afootx; avaxBeioriq; (Or. 5. 169cd). I do not see how this 
union could take place anywhere but in the world beyond. 340 

The union with the gods takes place in the imaginary world to which 
Attis and his followers gain access. Many scholars do not accept this 



338 See Turcan 1982a, 317: "Le salut des fideles est lie au salut du dieux dont ils 
partagent cultuellement les epreuves"; also J. Pepin, Reactions du christianisme latin a 
la soteriologie metroque. Firmicus Maternus, Ambrosiaster, S. Augustin, in Bianchi and 
Vermaseren 1982, 258—72 at 257—61. This text was referred to several times during 
the conference, e.g. by D.M. Cosi 1982, 489; Wedderburn 1982, 833 n. 66; R. Turcan 
1982b, 185 n. 9. In his defence of other-worldly salvation in the Phrygian cults, Cosi 
tried to contextualise it inter alia with a not less disputed phrase from CIL VI 510 = 
ILS 4152 — CCCA 3 no. 242: taurobolio criobolioque in aeternum renatus (376 CE), used by 
the parvenu Sextilius Agesilaus Aedesius to emphasise his rightful place among these 
Roman aristocrats (McLynn 1996, 327f). It is surely about time that the phrases used by 
the late-Roman elite ceased to be invoked as 'proof of wide-spread value-change. 

339 Turcan 1992a, 73f; cfl996b, 388f, 391f. Here again, the real difficulty is the 
nature of the literary sources. In the second-century Naassene 'hymns' I discussed 
above (n. 31), Attis appears as a god active in the world, anticipating what Julian says 
about him in Or. 5. Many critics however hold texts of this kind to be of no value in 
studying the true claims of the cult because of the contaminations they so evidently 
contain. The problem is methodological. Since they had no hierarchy with the author- 
ity to condemn their opponents, the oriental cults could not possibly develop the strict 
criteria of modern critics who are confident they can declare Julian atypical (not to say 
heterodox or heretical). Let me repeat: the oriental cults in the fourth century were 
indeed, in part at least, as Julian says they were; but in the second century they were 
different, since of course they changed over time. Turcan claims: "Les senateurs du 
Bas-Empire qui collectionnent ostensiblement titres sacerdotaux et initiatiques n'ont 
rien a voir avec le mithriacisme authentique de la grande epoque, celui des IP et IIP 
siecles apres J.-C": 1988, 260f What would we call the 'grande epoque' of Christianity 
then? I remind the reader of what I said earlier about ideas of purity and authenticity 
in areas such as these. 

340 I am glad to find D.M. Cosi likewise in opposition to the current minmalist trends: 
Salvatore et salvezza nei misteri di Attis, Aevum 50 (1976) 42—7 1 ; cf. idem 1986, 108H, 
where he links Julian's reinterpretation to sexual continence. 



138 CHAPTER TWO 

because they deny Attis' resurrection, some seeing it simply as a form 
of living death. 341 First of all, it is worth pointing out that the offer of 
resurrection need not only be made by deities that have themselves 
experienced death and resurrection: to deny the resurrection of Attis 
is not to show that his followers could therefore not have received any 
promise of a blessed existence in the world to come. That said, we may 
consider the terms of the denial. At one time people objected to the 
Frazerian category 'dying and rising gods'. 342 Others claim the idea of 
resurrection is simply a caique on Christian ideas. The real problem, 
to my mind, is to discover the true nature of the resurrection. Christ 
regained his human body, and the Church affirms belief not only in the 
eternal life of the soul but the resurrection of the body, on the model 
of Lazarus in the Gospel (John 11.1—46). According to the myth of 
Attis, however, his immortality was signalled by the perpetual wiggling 
of his little finger and the fact that his hair never stops growing. This 
is a completely different conception of immortality from belief in 
bodily resurrection. However, it is no less absurd to claim that Attis is 
just a zombie than it is to think that the myth deliberately formulated 
a charged symbolic image such as this merely in order to suggest that 
what awaits the initiate beyond the grave is an unreal existence, which 
therefore cannot be. Conceptually, no doubt, it is easier to imagine the 
world beyond with the same body as one had in life and in which one 
committed the sins that prevent the believer from proceeding straight to 
heaven and oblige him or her to spend time unwillingly in Purgatory. 
Among the other arguments is the absence of a cyclical repetition 
of Attis' resurrection (e.g. Sfameni Gasparro 1983, 227). It is difficult 
to understand this, for the repetition of the mystery of death and 
resurrection, even in Christianity, only occurs in a ritual context (quite 
unlike the cases of Prometheus or Sisyphus, for example). At the (later) 
Hilaria, the community of believers celebrated the god's resurrection, 
even if it was only expressed in such apparently trivial signs as his fin- 
ger and his hair, and they knew that the next year the festival would 
repeat his passion and resurrection, with its processions and dramatic 



341 E.g. Sfameni Gasparro 1983, 227. But note the fine remarks on the resurrection 
of Attis and its potentiality by Musso 1983, 141fE 

342 E.g. B. Alio, Les dieux sauveurs du paganisme greco-romain, RSPT 15 (1926) 
5—34; L. de Grandmaison, Dieux morts et ressucites, RcSR 17 (1927) 97—126; G.C. 
Ring, Christ's Resurrection and the Dying and Rising gods, CBQ6 (1944) 216—229; 
K. Priimm, I cosidetti 'dei morti e risorti' nell' Ellenismo, Gregorianum 39 (1958) 
411-39. 



SYSTEMS OF BELIEF 139 

performances that Christian writers so scorned. One can only say 
that modern scholars evidently have greater difficulties than believers 
in antiquity in accepting a resurrection whose function was to bring 
them close to god. In fact, a scholar like Pieter Lambrechts accepts 
that Antoninus Pius' reform (if that is what it was) extended not just 
to the rituals, such as the introduction of the taurobolium/ 'criobolium, and 
organisational details, such as the introduction of the Archigallus, but 
also to the belief-system, since it is only from that point on that Attis 
is to be found in the literary and epigraphic evidence, which, in his 
view, can hardly be a coincidence. 343 It is only after Antoninus Pius that 
Attis is found as a god of life and resurrection, calqued no doubt on 
the ideas of rebirth and renovatio current at the time, a context which 
well fits the introduction of a day, the Hilaria, to celebrate the god's 
resurrection (Macrobius, Sat. 1.21). 

However, the objections raised against regeneration in the Phrygian 
mysteries really stem from the quite inacceptable requirement that a 
myth be totally transparent to analysis. The vagaries of imaginative 
productions such as these inevitably mean that they resist straightforward 
interpretation. To resort to the same logic, the doubtful resurrection 
of Attis is the result of his doubtful death: we can say that this more 
or less dead zombie more or less comes alive again, as can be inferred 
from his apotheosis, through which he grants his followers a new order. 
The fact that there are no explicit literary texts about the hypothetical 
renewal of life procured by the god can be partially compensated by the 
archaeological evidence. On the interesting fourth- century silver salver 
in repousse work from Parabiago, for example, which I have already 
mentioned (PI. 5), Attis is seated next to Cybele in the conveyance 
drawn by lions, in what seems to me incontrovertibly an image of the 
hierogamic procession, with three Corybantes armed with dagger and 
shield dancing around them. 344 It is Attis, actually, who is holding the 



!l! Cf. Lambrechts 1962, 20—52, supported by Fishwick 1966. The Mainz defixiones 
now strongly suggest that Attis was a significant deity long before Antoninus Pius (see 
nn. 130, 131 above). 

314 T/he ear ]y scholarship dated it to II P , e.g. A. Levi 1935, 6—10 and C. Albizzati, 
La patera argentea di Parabiago, Athenaeum 15 (1937) 190, followed by Vermaseren 
1966, 27—30 [claiming that it is the lid of a box]. Most later commentators however 
rightly prefer the IV 1 ', e.g. K. Weitzmann, Age of Spirituality: Late Antique and Early Christian 
Art (New York and Princeton 1979) 185f. no. 164; L. Musso 1983, 106; H. Beck and 
P. Bol (eds.), Spdtantike undfriihes Christentum: Katalog (Frankfurt a.M. 1983) 531 fig. 138; 
J.M.C. Toynbee and K.S. Painter, Silver Picture-Plates of Late Antiquity: AD 300-700, 
Archaeologia 108 (1986) 15—66 at 29f; H.G. Gundel, ^odiakos: Tierkreisbilder im Altertum 



140 CHAPTER TWO 

reins, which, his supposed zombie-state notwithstanding, tells us that it 
is he who is directly in control of the entire scene. Its symbolism is very 
relevant here because, according to Julian, Attis is for ever tmovpycx; 
xfj Mr|xpi kcxI f|v{o%0(;, the servant and charioteer of the Mother, 'ever 
led upwards as though from our earth'. 345 The adventus of the divine 
pair is framed by heaven, represented by Helios and Selene, and earth, 
symbolised by the four groups of figures at the bottom. To the right is 
recumbent Tellus with the cornucopiae, accompanied by two Horai who 
are attracting her attention to the procession. In the centre are four more 
Erotes connoting the seasons, and below them two marine deities, pre- 
sumably Neptune and Thetis. On the left is another recumbent female 
figure holding a river-reed and lying on an amphora: she must recall 
the birth of Attis on the banks of the river Sangarius, whose daughter 
Nana conceived him after eating a nut from the almond-tree that grew 
from the blood that flowed from the torn-off testicles of Agdistis. On 
the centre-right of the dish, over against the divine pair, is a complex 
symbolic representation of Time. An Atlas emerges up to his waist 
out of the ground; on his head he wears, in the manner of Hercules, 
a cap made from a lion's mask. He is bearing the weight of a vast 
circle, which must represent the ecliptic, the apparent path of the sun 
around the earth, inside which there stands a young man, half-draped 



(Mainz 1992) 260 no. 190; Turcan 1992a, 74. I find the arguments of Musso quite 
convincing: "II tipo iconografico di Cibele sulla quadiga tirata da leoni in compagnia 
di Attis conosciuto unicamente attraversi le emissioni della prima serie di contorniati 
(tra il 365 e il 394), la presenza de Sol Invictus radiato su quadriga . . . secondo una 
'versione' iconografica favorita nel IV secolo ... la pertinenza dei confronti silistici istitu- 
ibili con altri argenti datati alia seconda meta del secolo (ad es. il missorio di Teodosio 
I) . . . concorrono a far ascrivere la fabbricazione del piatto di Parabiago ad eta non 
anteriore alia meta circa del IVsecolo d.C." (p. 106). The decisive indication is surely 
the gesture of Helios as he drives his chariot, which seems to be unknown before mid- 
IIP. I concede however that the issue of the date is not quite settled: Jackson 1994, 
153f; LeGlay 1981, 404 no. 20 ['date discutee '], and, more significantly, Eisner 1998, 
209 with fig. 136 [colour], all remain undecided. Vermaseren too: both in 1977, 69 
and 72f, and in CCCA 4, 107 no. 268, he simply declines to come down on one side 
or the other. Whether either Jackson or Eisner was aware of Musso 's work may be 
doubted, however; the others could not have been. As for the point about hierogamy 
cf. Levi 1935, 8; Musso 1983, 12. 

345 Julian, Or. 5. 171c (tr. W.C. Wright). As Turcan 1996b, 397f. remarks, the allu- 
sion to the chariot must evoke the myth of Plato's Phaedrus; on p. 403 he argues that 
for Julian, drawing on the Timaeus and Aristotelian cosmology, the individual soul is 
congruent with the World Soul, indeed they are in a sense consubstantial. 



SYSTEMS OF BELIEF 141 

and holding a sceptre, who is generally interpreted as Aion. 346 To the 
right of this group is a gnomon with a snake swarming up it. 

The salver, which measures 39 cms in diam. and weighs 3.55 kg, 
seems to have been found in a grave, where it had been placed over 
the mouth of a wine-amphora containing ashes. In the same tomb was 
found a silver-plated bronze knife (Levi 1935, no. 5). Taken together, 
the finds suggest that the owner of the tomb may have been himself a 
gallus, a self-castrated devotee of the goddess. That might also explain 
the use of imagery taken from other cults, if so it be. In my view, the 
fact that he had himself buried with such a precious object suggests 
that he believed in the omnipotence of the divine pair who are shown 
in an eternal process through the universe. Cosmic order replaces the 
chaos caused by Agdistis; Attis takes the reins of life after death for his 
followers who have passed through the initiatory tests, including castra- 
tion if need be, and are committed to following him like a cortege of 
Corybantes. 

At the end of this rather detailed discussion of the conception of 
life after death in the oriental cults, it should be clear that the differ- 
ences between scholars derive more from a priori considerations than 
a disinterested analysis of the evidence. It is however only human to 
think that the evidence fits one's own ideas better than those of other 
people. A further problem is the role of Christianity both in the selec- 
tion of the primary evidence and in the minds of scholars, so that 
the ancient debates re-echo, of course in different forms, in modern 
discussion. I would like to think that my arguments have shown that 
the oriental cults, as relatively autonomous religious formations, planted 
the issue of the world beyond in the collective imaginaire of the inhabit- 
ants of the Empire during the first four centuries of the Common Era. 
The surviving evidence does not of course allow us to reconstruct the 
vision of each mystery in detail, but we can assume that all gradually 
altered under the pressure of changing circumstances. Christianity too 
kept working away at its version of what happens after death without 
succeeding in giving it a fixed form until the Church established a 



346 LeGlay 1981, 404 no. 20, noting that his hand is near Aries, the first sign of 
the zodiac; Beck 1988, 56 n. 131. Apart from the items in LeGlay's bibliography, note 
G. Zuntz, Aion im Romerreich. Die archaologischen ^eugnisse. SB Akad. Heidelberg (Heidel- 
berg 1991) 35fE; G. Casadio, From Hellenistic Aion to Gnostic Aiones, in D. Zeller 
(ed.), Religion im Wandel der Kosmologien (Zurich 1999) 175—90; idem, s.v. Aion, in Jones 
2005, 1: 207-210. 



142 CHAPTER TWO 

hierarchical organisation capable of imposing a single authoritative 
account. Since nothing similar happened in the case of the oriental 
cults, we cannot construct a version for them that looks as finished as 
that of Christianity. 34 ' Nevertheless, there seems to me sufficiently hard 
evidence to allow us to make a start. In my view, those initiated into the 
oriental cults believed that, thanks to a once-and-for-all divine act, they 
could attain a blessed afterlife available only to those who participated 
in the rituals that provided the requisite knowledge, and provided that 
they lived in accordance with the ethical norms approved by the gods. 
We shall look at the content of these rules in the following chapter. 



3 " Note in this connection the interesting book by J. Amat, Songes et visions. L'au-deld 
dam la litterature latine tardive (Paris 1985) where the author looks at the construction 
of the Christian hereafter from the angle of the dream-work of the dead, the angels, 
demons, and journeys to the next world that give us details about the Paradise people 
longed for. If the oriental cults ever created such dense images, the relevant informa- 
tion has been almost wholly destroyed by those same Christians. On the arbitrary 
ways in which the information we do have has been used, mainly in relation to the 
cult of Isis, see Brenk 1993. 



CHAPTER THREE 
SYSTEMS OF VALUE 

Nobody sees Death, 
Nobody sees the face of Death, 
Nobody hears the voice of Death, 
Savage Death just cuts mankind down. 

Gilgamesh Xl.vi. tr. Dalley 1989, 108 

Sacrificium est . . . corpus etiam nostrum cum 
temperantia castigamus 

Augustine, De civ.Dei 10.6 1 

The aim of this chapter is to collect the scattered evidence for what 
might add up to the system of values of these cults. Those who believe 
that they constituted merely one modality within the larger pagan 
system will of course see no point in such a quest. 2 I am convinced 
however that, given the context of the Empire's market in religious 
options, they were a response to the anxieties generated by a shift in the 
religious paradigm, a shift whose underlying cause, as in the Cumon- 
tian grand narrative, was the incapacity of the politicised religion of 
Rome to meet the needs of the diverse population of a complex and 
multi-cultural empire. 

The oriental cults evidently did not offer their solution as an alter- 
native cultural model. Christianity on the other hand did gradually 
construct such a paradigm, which we tend to see as more finished 
right from the start. In fact, however, the process took centuries of 
struggle towards a Christian self-definition vis-a-vis the world, and of 



1 "Our body too [is] a sacrifice when we chasten it by temperance" (tr. P. Schaff). 

2 Emblematic of such a view are Erwin Rohde's remarks a propos the adherents 
of Dionysus: Psyche: Seelenkult und Unsterblichkeitsglaube der Griechen (Freiburg and Leipzig 
1894 (1898 2 , repr. Darmstadt 1961) 391. He claims that once the ceremonies were 
over no profound trace of them remained in the hearts of these ecstatics, who received 
no impulsion to alter their conduct, no change of heart. He surely thought the same 
about all the other mysteries. Generally relevant here is the ancient construction of 
highly-coloured moral invective against opposing individuals and religious groups, esp. 
mysteries, which would have been pointless if the latter had had no ethical norms, 
explicit or implicit: R.M. Grant, Charges of Immorality against Various Religious 
Groups in Antiquity, in Van den Broek and Vermaseren 1981, 161—70. 



144 CHAPTER THREE 

internal confusion, until a uniform pattern emerged. 3 The impression 
of coherence is the result of a process of intellectual and physical 
cleansing that has removed the sophistries and blood from the history 
of the Early Church. Such a process was quite foreign to the oriental 
cults, although they did experience changes both in their systems of 
belief, as we have seen, and in their rituals, which we shall look at in 
the next chapter; and their values too must surely have adjusted slowly 
to historical change. Unfortunately the evidence is simply inadequate 
to allow us any insight into such shifts over time, and we must be con- 
tent with the more modest task of filling in the main outlines of their 
ethics within the wider cultural context, and considering how their 
adherents responded to divine injunctions. The aim is thus to provide 
a preliminary sketch-map of morality in these cults. 

On the authority of Cicero, who translated Greek f|9iKO<; as moralis, 
we ordinarily treat morality and ethics as synonymous terms. In some 
contexts, however, the words need to be differentiated. Foucault for 
example has this to say about the word morality: 

By 'morality' one means a set of rules and values through the intermedi- 
ary of various prescriptive agencies such as the family (in one of its roles), 
educational institutions, churches and so forth. It is sometimes the case 
that these rules and values are plainly set forth in a coherent doctrine and 
an explicit teaching. But it also happens that they are transmitted in a 
diffuse manner, so that, far from constituting a systematic ensemble, they 
form a complex interplay of elements that counterbalance and correct one 
another, and cancel each other out on certain points, thus providing for 
compromises or loopholes. With these qualifications taken into account, 
we can call this prescriptive ensemble a 'moral code'. But 'morality' also 
refers to the real behaviour of individuals in relation to the rules and 
values that are recommended to them: the word thus designates the man- 
ner in which they comply more or less fully with a standard of conduct; 
the manner in which they obey or resist an interdiction or prescription; 
the manner in which they respect or disregard a set of values. 4 



3 Cf. the excellent account of Lieu 2004. 

1 Foucault 1985, 25. Foucault's work must be approached with caution, as the 
feminists have shown. But it does at least set up a model that can be criticised, cf. the 
abundant information, debate and bibliographies in e.g. Goldhill f995; D.HJ. Lar- 
mour, P.A. Miller and C. Platter (eds.), Rethinking Sexuality: Foucault and Classical Antiquity 
(Princeton 1998). On a quite different tack, J. Annas, The Morality of Happiness (Oxford 
and New York 1993) offers a densely-written review of the relevant currents of Greek 
thought and its influence on modern morals and ethics, with discussion of many points 
relevant to my project here. 



SYSTEMS OF VALUE 145 

Ethics however is a practical attitude of mind committed to construct- 
ing a coherent pattern out of various existential values. 1 Such an effort 
will obviously be heavily determined by the cultural attitudes that 
influence behaviour at the level of attitude rather than action. Actions 
will of course be prefigured by attitudes in such a way that "knowing 
'how to act' as a member of a community in the ancient world . . . was 
inseparable (with the exception of an insignificant number of indi- 
viduals) from the practice-and-belief of religion" (Gordon 1979a, 19). 
Religion provides a set of moral values in keeping with the surrounding 
social conditions, thus predisposing individuals to behave in accordance 
with a particular set of ethical values. 6 There is however yet a third 
aspect here, which I would have liked to examine if the evidence had 
been there, namely the role of the individual in constructing ethical 
or moral conduct (here the terms can again be used interchangeably). 
As Foucault says: 

A rule of conduct is one thing: the conduct that may be measured by 
this rule is another. But another thing still is the manner in which one 
ought to 'conduct oneself — that is, the manner in which one ought to 
form oneself as an ethical subject acting in reference to the prescriptive 
elements that make up the code . . . The determination of the ethical substance 
[. . .] is [. .] the way in which the individual has to constitute this or that 
part of himself as the prime material of his moral conduct . . . There are 
also possible differences in the forms of elaboration, of ethical work {travail 
ethique), that one performs on oneself, not only in order to bring one's 
conduct into compliance with a given rule, but to attempt to transform 
oneself into the ethical subject of one's own behaviour.' 



5 Cf. J. Vidal, Rite et ritualite, in Ries 1986b, 53. T. Rendtorff puts it succinctly: "Eth- 
ics is the theory of the conduct of human life": Ethics. 1: Basic Elements and Methodology 
in Ethical Theology; 2: Applications of an Ethical Theology. Transl. K. Crim (Philadelphia 
c. 1986-89) 1: 3. 

6 Cf. W.A. Meeks' formulation: "Morality names a dimension of life, a pervasive 
and, often, only pardy conscious set of value-laden dispositions, inclinations, attitudes, 
and habits" (1993, 4). 

' 1985, 26f. emphases in the original. His view here seems to me close to that of 
Meeks: "I take 'ethics' in the sense of a reflective, second-order activity: it is morality 
rendered self-conscious; it asks about the logic of moral discourse and action, about 
the grounds for judgement, about the anatomy of duty or the roots and structure of 
virtue . . ." (1993, 4). However one also needs to bear in mind the more restrictive view 
described (but not shared) by Hans-Dieter Betz: "Ethics as a term tends to be defined 
either on the model of modern ethical theory as a general and scientific theory of 
moral obligations, or on the model of classical philosophical ethics as a rational system 
of moral virtues necessary to bring raw human nature up to the level of eudaimonia. If 
these are the only admissible definitions of ethics, primitive Christianity did not and 
could not have had such ethics" (1978, 2fi). 



146 CHAPTER THREE 

The voluntary decision to join an existing group of initiates that, as we 
shall see, possessed a set of moral rules in some tension with the wider 
cultural context presupposes that the individual had recourse to this 
third level, necessitating the development of the notion of a conscience, 
and thus making him personally responsible for his actions. Initiation 
was understood not as a nice little ritual but as the expression of a 
deep religious conviction. This surely satisfies the conditions for us to 
be able to speak of morality, ethics and values in the wider sense in 
the context of these cults. However what we have here is not an ethos, 
understood as a set of beliefs and attitudes characteristic of a social 
group. It is neither a set of maxims derived from 'folk morality' nor 
a psychotherapeutic vademecum developed by some philosopher (cf. 
Betz 1978, 2f). It is rather a complex and dynamic construction that 
drew on both popular ethics and philosophical injunctions, and thus 
created its own rich moral world. 

But of course the oriental cults did not exist in a vacuum. If they were 
able to offer alternatives to the dominant paradigm, it is because there 
was a demand for them. From the late Hellenistic period, but mainly 
in the Principate, some 'sacred laws', to say nothing of the famous text 
at Epidaurus, and the Hippocratic Oath, required not merely physical 
purity but also an undefiled mind. 8 Ritual lustration was increasingly 
taken to connote, and presuppose, moral integrity 9 The Samothracian 
mysteries had the reputation of making their initiates e\)oe(3eoxepov)(; 
Kai SiKouoxepoiK; koc! Katoc nav peA/uovaq (Diod. Sic. 5.49.6). 10 One 
of the best-attested cases is the well-known prescriptions of a private 
cult-association at Philadelphia in Lydia, dated to the late second, or 
early first centuries BC. 11 Dionysios, the leader of the group, claims 



8 Cf. J.N. Bremmer, How Old is the Idea of Holiness (of Mind) in the Epidaurian 
Temple Inscription and the Hippocratic Oath?, %PE 142 (2002) 105-07, citing nine 
parallel formulations, the earliest from the first century CE. Bremmer shows that the 
traditional fourth-century BC date for the Oath and the temple-inscription cannot 
be correct. 

9 A. Chaniotis, Reinheit des Korpers — Reinheit des Sinnes in den griechischen 
Kultgesetzen, in J. Assmann and T. Sundermeier (eds.), Schuld, Gewissen und Person 
(Gutersloh 1997) 142-79. 

10 Cf. S. Guettel Cole, The Mysteries of Samothrace during the Roman Period, 
ANRWII.18.2 (1989) 1564-98 at 1577; and eadem 1984, 31, critical of Karl Lehmann's 
claim that there was a confession of sins in this cult before £7t6jtT£ia. 

11 Syll. 985 = E Sokolowski, Lois sacrees d'Asie mineure (Paris 1955) 53—58 no. 20, 
with the now somewhat out-dated commentary by Barton and Horsley 1981, itself 
complementary to the study of the gods listed in the inscription by O Weinreich, 
Stiftung und Kultsatzungen eines Privatheiligtums in Philadelphia in Lydien, Sitzungs- 



SYSTEMS OF VALUE 147 

as his authority dreams sent to him by Zeus (11. 4 and 12f.), who was 
apparently the main divinity worshipped. 12 The rules themselves, mainly 
concerned with the interdiction of sexual contact outwith marriage, but 
also of abortifacients and love-charms, bear a family resemblance to the 
so-called 'confession texts' from the same general area. 13 Members seem 
to have been required to swear an oath to abide by them. In case of 
disobedience, offenders are threatened with divine hatred and dreadful 
punishment (ueydXat; . . . xiucopiocc; 7iepi9r|oo\)[oiv: 1. 49f). 

The text's main interest in the present connection, however, is the 
insistence on purity of heart and mind, as a duty and service to the 
gods. The goddess Agdistis, the moral guardian of the association, is 
asked to engender [dycxBac;] Sicxvoicxc; avSpdai Kcxi ywai^iv [eA,ei)9epoic; 
Kai] SouA-ok; (11. 52-4), the good disposition and moral judgement that 
makes it easy to keep the rules. This innocence of mind is to be made 
visible to all at the monthly meetings for sacrifice, when the members 
are required physically to touch the inscription to show that they have 
kept the rules (11. 54-60). 14 This practice is evidently modelled on the 
local form of ordeal: those who perjured themselves were inviting the 
gods to punish them; those who refused to touch the inscription were 
as good as confessing that they had failed to keep to the rules. 

During the same period, philosophy was occupied in creating a body 
of reflection on such matters. A sententia by Seneca hints at the way 
an individual, even a slave, should involve himself in the creation of 
a personal moral character, as a sort of dynamic assault on destiny: 



berichte der Akademie der Wissenschaften in Heidelberg, phil-hist. Klasse 1919 no. 8 (Heidelberg 
1919) 1-68. 

12 The word used for the group is oikoi^ which at the same time seems to refer 
to a room in Dionysios' house, where the group met: Barton and Horsley 1981, 15f. 
The association was open to men and women, slave and free (Borgen 1994, 58). 
Perhaps for that reason particular attention had to be paid to sexual purity (cf. 86^r|i 
Trjt apiaiTji, 1.2). 

13 G. Petzl, Die Beichtinschriften im romischen Kleinasien und der Fromme und 
Gerechte Gott, Westfalische Akademie der Wissenschaften, Vortrdge G355 (Opladen 1998); 
A. Rostad, Confession or Reconciliation? The Narrative Structure of the Lydian and 
Phrygian 'Confession Inscriptions', SymbOslo 11 (2002) 145—64; A. Chaniotis, Under 
the Watchful Eyes of the Gods: Divine Justice in Hellenistic and Roman Asia Minor, 
in S. Colvin (ed.), The Greco-Roman East (— Tale Classical Studies 31) (Cambridge 2004) 
1—43; A. Rostad, Human Transgression — Divine Retribution, diss. Bergen 2006. Remarkably 
enough, Barton and Horsley fail to make this connection. 

11 Barton and Horsley 1981, 14 n. 30 also stress the unusual expression used in 
this connection: ooot Jtiaxeioovoiv eat)ToT<; (1. 56), for which they prefer the translation 
'those who have confidence in themselves'. 



148 CHAPTER THREE 

Sibi quisque dat mores, ministeria casus adsignat. 15 The anti-Christian pub- 
licist Celsus unfavourably compares Jesus' announcement that sinners 
were welcome in the Kingdom of God (Matth. 7.7; Luke 19.6) to the 
practice of the mysteries: 

oi |iev yap ei<; xaq aXkaq xekexaq Kakovvxeq TipoKrip-uttouci tdSe- oaxiq 
Xeipa<; Ka9apo<; Kai (pcovqv avvexbq, Kal avQiq etepor oaxiq ayvbq anb 
navxbq txvaovq Kal otco r\ yuxri o\)8ev cruvoi8e kokov, Kal oxcp ev Kal 
8iKaico<; (kpHcorai. 16 

We find similar sentiments being adduced nearly two centuries earlier 
by Philo, who claims that it is unthinkable that the truly sacred should 
be defiled with avtipoic; Qvaiaic,, i.e. sacrificial offerings made by those 
who are have sinned." It has been argued that such views are in fact 
old, and go back ultimately to the formulations and injunctions of 
sacred laws, perhaps even in Archaic times. 18 

Since the members of these initiatory communities lived in the wider 
society, it was the cultural values of the latter that to a large extent 
determined their ethical demands. 19 Indeed, it must be granted that the 
moral system propounded by these cults was closely bound up with the 
general morality of the society they hoped to succeed in. The aim of 
initiation was not to cut the individual off from social life or to turn 
him into an outsider, a sort of Timon of Athens. Quite the reverse: 
their aim was integration, since otherwise they would not merely enjoy 



15 "Each individual acquires his moral character for himself: chance allots his duties": 
Seneca, Epist. mor. 47.15, in the context of a discussion about relationships between 
master and slave; the previous sentence reads: non ministeriis illos aestimabo, sed moribus; cf. 

0. Tescari, Echi di Seneca nel pensiero cristiano e viceversa, Unitas 2 (1947) 171-81; 

1. Hadot, Seneca und die griechisch-mmische Tradition der Seelenleitung (Berlin 1969); U. Domin- 
guez del Val, El senequismo de Lactancio, Helmantica 23 (1972) 291-323. 

16 "Those who summoned persons to participate in mystery-rites proclaim the fol- 
lowing: Whoever has pure hands and who speaks intelligibly (i.e. is of sound mind; 
or perhaps: speaks Greek); yet others utter this proclamation: Whoever is pure of all 
taint and whose mind is conscious of no ill and who has lived a good and righteous 
life": ap. Origen, Contra Celmm 3.59; cf. Julian, Or. 7.29, 239c. 

17 Spec. leg. 1.270 discussing Numbers 19.1-10. 

18 See the discussion by Dickie 2004, 579-90. 

19 I use the word 'community' with some reservation in this context, since it is 
extremely difficult to say much about either the structure or the permanence of groups 
such as the dendrophori or the pastophori, a task I do not attempt here. I have deliberately 
preferred to present a rather idealised picture of associations founded on common 
values. On the difficulty of defining religious communities in antiquity, see the useful 
comments of Belayche 2003; cf. North 2003, stressing the role of personal choice, and, 
by implication, loose affiliation. 



SYSTEMS OF VALUE 149 

no cultural influence but would end up by dying out. 20 They might be 
superficially different from the dominant paradigm, but ultimately their 
content was determined by the historical context. 21 

Since neither a discussion of such matters nor any general statement 
by initiates survives from antiquity we are not in a position to decide 
whether the moral system in each case (itself a creation over time) bor- 
rowed from an independent or prior system. In my view, the individual 
myths had their own moral content, which could be adapted in differ- 
ent ways in different periods depending on the historical circumstances; 
it is this content that the Christian apologists attack. However I also 
believe that the values I postulate altered as the oriental cults shrugged 
off their image as exotic outsiders and became a means of transform- 
ing paganism. I shall be looking at this in due course, insofar as the 
surviving evidence permits. 

It is not my intention here to examine the principles of the Roman 
moral system, the value-system to which these cults had to adapt them- 
selves (even if they did develop their own ethics in some areas). Accord- 
ing to Mario Vegetti, "the impromptu quality, socially and culturally 
speaking, of the processes of being brought up and becoming a moral 
subject in antiquity meant that there was plenty of opportunity for 
uncertainty and conflict, and thus of choice and freedom unknown in 
other social systems". 22 In fact this freedom applied mainly to the realm 



20 We may here observe that it was the temple of Isis and not any of the prominent 
temples in the Forum of Pompeii that were restored after the devastating earthquake 
of AD 62. CIL X 846 = ILS 6367 records that the decuriones were glad enough to 
elect the donor's son N. Popidius Celsinus to a seat on the council despite his tender 
age (6 years). "The honor also shows the political influence wielded by devotees of 
Isis in the town at this date": Zanker 1998, 126f. Much the same can be said about 
the array of Isiac statues at Athens from c. AD 50 until the end of II P : Walters 1988, 
59 with Appdx. Ill; 2000, 87-9. 

21 This must be the explanation for the similarity between Plutarch and Christian 
ethics that so surprises some people that they take to postulating an improbable and 
undemonstrable Christian influence on the Chaeronean sage. It would in fact be far 
more sensible to assume that the Christian writers were drawing on the 'folk morality' 
available to them, which itself derived from the oral dissemination of philosophical 
teachings (Betz 1978, 8). I myself think that the philosophers were often simply giving 
a special twist to a common fund of moral sentiments shared by themselves and their 
contemporaries, pagan and Christian. Wolf Liebeschuetz, on the other hand, has well 
examined the subtle links between Roman religion and its moral implications, which 
were not understood by most people (1979, 39-54). 

22 L'etica degli antichi (Rome and Bari 1989) 5. On the basis of an impressive amount 
of evidence, this book argues the thesis that the true constants over the longue duree 
are happiness and virtue. He deals with the oriental cults along with a series of 



150 CHAPTER THREE 

of the imagination, since the rules ended up by being quite restrictive, if 
not perhaps as restrictive as in a society where morality was in general 
stricter. 23 It is anyway clear that societies are not equally repressive all 
down the line: in many ways, though not all, the rules are tighter for 
those who do not enjoy the privilege of belonging to the hegemonic 
order. 24 Everywhere however there are escape-mechanisms, that is, 
supplementary or sub-cultural value-systems, which help to sublimate 
the differences and allow the individual a degree of self-awareness in 
liminal areas. They are escape-routes that make it possible to bear liv- 
ing in a stiffingly enclosed world by offering alternatives to the model 
set up by the cultural norm. 

These contrasting but complementary systems created a good deal 
of ambiguity, which makes interpretation still more difficult. 25 Where 
our information is more abundant, the ambivalent metaphors may 
sometimes allow us to gain some idea of what is merely tacit, what does 
not need to be, or cannot be, said; but when every scrap of evidence 
is scrutinised under a microscope for its quantum of truth, as in the 



alternatives to the traditional system on pp. 304ff, though without any special points 
relevant to my purpose here. 

23 J.G. Griffiths, The Divine Verdict. A Study of Divine Judgement in the Ancient Religions. 
Suppl. Numen 52 (Leyden 1991) 45—109 argues in favour of the idea of a Greek moral 
law (and, mutatis mutandis, a Roman law) with divine sanction. This is a completely 
different way of conceptualising the problem of morality and its reinforcement from 
that adopted by the work cited in n. 27. Griffiths' style of argument is admittedly 
old-fashioned, but we can agree that explanations of human misfortunes, especially 
illness, poor harvests, famine and so on were regularly couched in religious terms, as 
punishments for individual or collective wrong-doing, cf. recently G.E.R. Lloyd, In the 
Grip of Disease: Studies in the Greek Imagination (Oxford 2003); and we shall have occa- 
sion below to note one or two instances in relation to Isis. Since health and prosperity 
were the goods regularly requested from the divine world, and it was a foundational 
belief that piety led to good fortune, it is inevitable that their being withheld should 
be considered a result of impiety, i.e. some sort of wrong-doing. The general function 
of the oracular system was to provide the information required, at every level, to sort 
out such muddles. 

24 On the difference between Roman and modern conceptions of 'public' and 'pri- 
vate', see K. Cooper, Closely Watched Households: Visibility, Exposure and Private 
Power in the Roman Domus, Past & Present 197 (2007) 3—33. 

25 The introduction of new cults into Rome and the polymorphism of its relations 
with the divine obviously increased the range of religious modes and experiences 
available. This was not so much a matter of individual choice (which seems to me a 
typical liberal myth) as of the mobility of deities facilitating their claim to universal 
authority and range. But see Hopkins 1999, 82. 



SYSTEMS OF VALUE 151 

case of the oriental cults, there is hardly room for such non-verbal 
perception, social or cultural. 26 

It is also not easy to say precisely what the dominant cultural norm 
was. According to L.R. Lind the ten basic ethical values, at least as 
regards Roman morality, were fides, qffwium, prudentia, constantia, utilitas, 
diligentia, religio, pietas, fortitudo, virtus. These notions were in their various 
ways deeply embedded in social praxis. 27 One or two comments are 
called for here. These values were in fact those of an upper-class male 
Roman. 28 That is surely why there is no specific reference to sexual 
morality, which, by contrast, is one of the best-documented areas in 
the moral thinking not only of the Romans but also of all cultures (and 
religions). 29 In order to tackle the concept of virilitas one would naturally 



26 The situation in the case of early Christianities is quite different. I much admire 
the way Meeks 1993, 37ff deals with the problem. On p. 216, he brilliantly defines 
Christian ethics as a polyphony, by which he means the apparendy contradictory values 
developed by (Christian) communities over time, f see this less as a principle than as 
an inevitable and universal consequence of historical change, as I shall try to establish 
in the course of this chapter. 

2/ L.R. Lind, The Idea of the Republic and the Foundations of Roman Morality: 
1, in C. Deroux (ed.), Studies in Latin Literature and Roman History, 5. Collection Latomus 
206 (Brussels 1989) 5-34; idem, ~ 2, in C. Deroux (ed.), Studies in Latin Literature and 
Roman History, 6. Collection Latomus 217 (Brussels, 1992) 5-40. C. Skidmore, Practical 
Ethics for Roman Gentlemen. The Work of Valerius Maximus (Exeter 1996) 53-84, extracts 
from Valerius Maximus the following nine virtues: virtus, moderatio, continentia, verecundia, 
severitas, pietas, aequitas, constantia, gratitudo. See also G.G. Belloni, Note sulle virtu romane, 
Aevum Antiquum 1 (1988) 181-192, who picks out a rather narrow group, based on the 
list of those that were divinised and actually worshipped: lihertas, ius and iustitia, aequi- 
tas, fides; A. Michel, La vertu n'est-elle qu'un mot? in P. Dumont (ed.), Problemes de la 
morale antique (Amiens 1993) 123-31. Other 'virtues' have been studied from a variety 
of points of view, e.g. H. Wagenvoort, Pietas, in idem, Pietas: Selected Studies in Roman 
Religion (Leyden 1980) 1-20; H.I. Marrou, L'ideal de la virginite et la condition de lafemme 
dans la civilisation antique (Paris 1953) (both typical of their time); virtues in the ancient 
novel: G. Schmeling, Manners and Morality in Apollonius of Tyre, in P. Liviabella Furiani 
and A. Scarcella (eds.), Piccolo mondo antico (Perugia 1989) 197-215. It is often thought 
that there was a sharp difference between the morality of Early Christianity and the 
pagan environment. Meeks 1993, 68f£ argues, quite to the contrary, that there was in 
fact a considerable overlap between some pagans and some Christians. Here, obvi- 
ously, it is all a matter of which texts one cares to use, that is, of the subjective views 
of the historian. I think Keith Hopkins and I share a good deal of common ground 
here (1999, 201 n. 46). 

28 Cf. now R.M. Rosen and I. Sluiter (eds.), Andreia: Studies in Manliness and Cour- 
age in Classical Antiquity. Penn-Leiden Colloquium on Ancient Values, 1 . Mnemosyne 
Suppl. 238 (Leyden 2003); M. McDonnell, Roman Manliness: Virtus and the Roman Republic 
(Cambridge 2006). 

29 Think of Martial, for example, whose jokes allow us to trace what at first sight 
seems a fuzzy line between good and evil in his ethical perspective, well discussed in 
a series of publications by Marguerite Garrido-Hory La vision du dependant chez 



152 CHAPTER THREE 

have to look at the discourse of the body and the very formal nature 
of sexual relations, both of them areas that were avoided in the older 
literature. 30 Attitudes to what Foucault calls the aphrodisia, and other 
aspects of the notion of virilitas, naturally shifted over a man's lifetime. 
By contrast, the virtues connected with the realm of sexual morality, 
among them pudicitia, castitas, pudor, concordia in marriage (one of the 
meanings of the dextrarum iunctio at the marriage ceremony), and even 
amor affected women directiy. All this is clearly expressed in the literary 
and epigraphic evidence for laudationes and consolationes. 31 

Foucault has claimed moral thinking about pleasure in antiquity 
neither developed nor tried to develop a codification of acts, nor was 
there any hermeneutics of the subject in this area. What it did do was 
to stylise an existential aesthetic: 

The relation to truth [in Platonic discourse] was a structural, instrumental 
and ontological condition for establishing the individual as a moder- 
ate subject leading a life of moderation; it was not an epistemological 



Martial a travers les relations sexuelles, Actes du XI' collogue du GIREA, Kazimierz 1980, 
in Index 10 (1981) 298-315; La femme chez Martial, Hommages R. Fietier, Paris, 1984, 
301-31 1; Juvenal: Enclaves et ajfranchis a Rome (Paris 1998); Femmes, femmes-esclaves et 
processus de feminisation dans les oeuvres de Martial et de Juvenal, in AA.VV. 1999, 
303—314. There are however several different ways of studying Roman sexuality, cf. 
the excellent collection by Hallett and Skinner 1997; R. Langland, Sexual Morality in 
Ancient Rome (Cambridge 2005). 

30 Cf. J. Walters, Invading the Roman Body: Manliness and Impenetrability in Roman 
Thought, in Hallett and Skinner 1997, 29—43, who comes to the same conclusions as 
Garrido-Hory in the 1980s. Much the same can be said about R.A. Pitcher, Martial 
and Roman Sexuality, in T.W. Hilliard et al. (eds.), Ancient History in a Modern University, 
1 (Cambridge 1998) 309-15, written in complete ignorance of the work of Garrido- 
Hory, and whose results are less complete and less systematic than hers. 

31 Cf. the Augustan texts known as the laudatio Murdiae (CIL VI 10230 = IIS 8394) 
and ~ 'Turdiae' (CIIVI 1517 = ILS 8393), cf. W. Kierdorf, Laudatio funebris. Interpretationen 
und Untersuchungen zur Entwicklung der romischen Leichenrede (Meisenheim am Glan 1980, 
with an important section on virtus; F. Gasco, in L. Garcia Moreno, E Gasco, J. Alvar 
and FJ. Lomas, Historia del mundo cldsico a braves de sus textos, 2 (Madrid 1999) 129ff As for 
consolationes, note esp. the texts of Seneca and Plutarch, and the consolatio ad Liviam: ed. 
H. Schoonhoven, The pseudo-Ovidian Ad liviam de morte (Groningen 1992). All stress the 
range of duties open to a woman in the Empire and are good sources for studying my 
theme here. On a related topic, the conceptualisation of marriage and the social role 
of women, see G. Williams, Some Aspects of Marriage Ceremonies and Ideals, JRS 
48 (1958) 16-29; J. Gage, Matronalia. Collection Latomus 60 (Brussels 1963) 100-53; 
K.R. Bradley, Ideals of Marriage in Suetonius' Caesares, Rivista Storica dellAntichita 15 
(1985) 77-95; Grubbs 1994 (cf. p. 404 n. 60 below); R.D. Brown, Lucretius on Love and 
Sex: A Commentary on RN IV1030—1287. Columbia Studies in the Classical Tradition 
15 (Leyden 1987) 122—27. On the expression of love in economic terms through testa- 
ments: E. Champlin, Final Judgements: Duty and Emotion in Roman Wills, 200 BC—AD 250 
(Berkeley and Los Angeles 1991). 



SYSTEMS OF VALUE 153 

condition enabling the individual to recognize himself in his singularity 
as a desiring subject and to purify himself of the desire that was thus 
brought to light. 

This is schematic, but will do. By contrast in Christian ethics great sig- 
nificance was attached to "two opposite yet complementary practices: a 
codification of sexual acts that would become more and more specific, 
and the development of a hermeneutics of desire together with its 
procedures of self-decipherment". 32 This is Foucault's grand narrative 
in The History of Sexuality, vols. 2~ 3. 

On the other hand, Lind's rather academic list is very selective. They 
are taken exclusively from a specific type of literary production. What 
does it mean to claim that these were the main or central values of 
the Roman citizen? They really amount simply to Valerius Maximus' 
affirmation of the paradigm of good-natured seriousness on the part of 
the dominant groups. 33 Moreover the terms are understood in a purely 
lexical fashion and assume extensive translatability. There is little sense 
of the dynamics of such concepts and their role in social negotiation 
and the maintenance of 'face'. 34 Moreover we need to distinguish 
between what Robert Kaster terms 'dispositional' and 'occurrent' uses 
of behavioural words. 35 Getting at the values of the dominated is still 
more tricky, because there is no point in projecting onto the entire 
population of the Empire the ideologically-driven moral aspirations 
of the dominant class. It is to be hoped that these silenced voices may 
gradually be recuperated by dint of patient analysis of sayings, refrains, 
fables and graffiti: such work will enable us to describe attitudes that 
were often enough quite different from those that feature in accounts 



32 Foucault 1985, 89 and 92. 

33 Cf. W.M. Bloomer, Valerius Maximus and the Rhetoric of the New Nobility (Chapel 
Hill 1992) 22-25,184-229; P. Combes, Valere Maxime, Lions 1 111 (Bude) (Paris 1995) 
20-46. 

34 Note A.M. Eriksen, Redefining Virtus: The Settings of Virtue in the Works of 
Velleius Paterculus and Lucan, in Ostenfeld 2002, 1 1 1—22; cf. Pliny the Younger's 
account of honestas: N. Methy Les lettres de Pline le Jeune: Une representation de t'homme 
(Paris 2007). 

35 Kaster 2005, 24—27, though terms need to be transposed slightly for 'virtues'. He 
defines 'dispositional' as a general sensitivity to 'appropriate' situations, and 'occurrent' 
as particular experiences in particular situations. The book's stress on the dynamics of 
emotions and the relevant 'scripts' could rewardingly be applied to virtues. Note esp. 
the excellent analyses of pudor, understood as an emotion (ibid. 28—65), and integritas 
(134-48). 



1 54 CHAPTER THREE 

of the moral thinking of the elite. 36 Nevertheless, one has to start some- 
where, and for all its abstraction Lind's list is a useful aide-memoire of 
the topics I shall be discussing in this chapter. 

1 . Between Utopia and Reality 

o toi 7itepoei(; eppiye Ylayaaoq 
Seanoxav eSeXovt' eq oupavou araGiaoxx; 
eA,0eiv |ie0' 6|idyupiv BeXXepcxpovxav 
Zr\voq 

Pindar, Isthmians 7.4 z f-47 37 

As will now be clear, I believe the oriental cults developed their own 
ethical norms, though for lack of documentation we are hardly in a 
position to recuperate them. 38 One or two scraps of information how- 
ever do provide grounds for the claim that they developed their own 
ethical models, quite aside from purely aesthetic matters. 39 In the case 
of the cult of Mithras, there is a fragmentary line from the mithraeum 
of S. Prisca on the Aventine: . . .perlata humeris [- -J mfajxima divum, which 
must mean something like '(bear) the commands of the gods on (my or 
your) shoulders to the very end'. 40 Whatever the precise meaning, the 



36 For one promising approach, see the paper by the late J. Cascajero, Historia 
Antigua y fuentes orales, Gerion 17 (1999) 13—57, with the earlier bibliography. 

37 "Indeed, winged Pegasus threw his master, when Bellerophon desired to enter the 
habitations of heaven and the company of Zeus" (tr. W.H. Race). 

38 Cf. Pleket 1981; cf. Henrichs 1984. 

39 R. Rubio, La propaganda de la estetica: simbolos exoticos del individuo en la 
difusion de los misterios orientales, in J. Alvar, C. Blanquez and C.G. Wagner, (eds.), 
Formas de difusion de la religiones antiguas. ARYS 3 (Madrid) 219-30. 

40 Vermaseren and van Essen 1965, 204f. no. 9 = Sanzi 2003, 439: Mithras no. 26.1. 
The reading is, as so often with Vermaseren, conjectural. His reading of the complete 
line is: Atque perlat(a) humeris tfujli mfajxima divum. According to V, this is a pentameter, 
which is impossible. For it to be a perfect metrical hexameter, as it otherwise appears 
to be, we would need three long syllables in the fourth foot, for which there is no 
space — indeed even his tfujli seems to me quite imaginary. To my mind, we should 
assume 1) a short word here such as tibi (i.e. at Mithras' command), and 2) that there 
were two feet before perlata (the reading atque is very uncertain). As for the word maxima, 
V concluded that it meant 'commands', whereas Sanzi translates literally 'le cose piu 
importante'. It is unclear whether the sentence is a first-person utterance by Mithras 
or by a Mithraist; either way, it would be unique among the lines at S. Prisca. If it is 
an utterance by Mithras, the primary reference would be to carrying the bull on his 
shoulders before killing it. This would confirm Cumont's belief that Mithras received 
a command to kill the bull. If it is an utterance by an adherent, identifying himself 
with the god, the metaphorical meaning would be primary. 



SYSTEMS OF VALUE 155 

implication is that the source of morality is the divine will, and that it 
is the individual's moral duty to act in accordance with this will. Then 
there is what Hermes is made to say to Julian at the end of the Caesares: 
SeScdkcx xov naxzpa M(9pocv zmyv&vca- av Se ocutot) tcov evtoXcov 
e^oi) . . ., 'I have granted (you) to know Mithras the Father: (it is your task 
to) keep his commands in mind' (336c). The word evxoA,r| also occurs 
in Christian moral contexts implying the demand for moral purity 41 In 
my view, it is irrelevant here that Julian uses the name Mithras fairly 
systematically also to refer to his syncretistic solar deity: this deity was 
a composite of various other solar deities, so that the cult of Mithras 
may well have contributed this item to the wider concept. 42 

We can also, I think, adduce the fact that the oriental cults were 
sometimes mocked in performances of mimes, and in festival proces- 
sions. 43 Outsiders laughed at their gods, the very basis of the norms 
governing their conduct; and at the very fact that they were looking 
for alternatives to standard rules. Ordinary people thus experienced a 
brief sense of liberation in making fun, in the presence of believers, of 
what the latter felt to be sacrosanct. In such cases, laughter is mainly 
evoked by what is, culturally speaking, experienced as oppressive. Since 
the adherents of the oriental cults did not find their gods oppressive, 
they of course did not find the mockery funny. We have no evidence of 
initiates' reaction, but there is a passage from Clement of Alexandria 
that may give us a hint: 

For since all forms of speech flow from mind and manners, there would 
be no mockery if there were no mocking mimes ... So if comedy (love, 
yeX(x>xonoioi(;) is to be ejected from our society, we ourselves ought not 
to stir up laughter. For it would be absurd to imitate things we are not 
allowed to hear; and still more absurd for a man to set about making 
himself a laughing-stock, that is, the butt of insult and derision. For if 
we could not endure to make ourselves ridiculous, as we see some do 
in processions (ev xaic, 7to|i7taic,), how could we bear to have the true, 



41 E.g. Justin, \Apol. 16.6; 66. The word is very common in NT: W.F. Moulton and 
A.S. Gedes, A Concordance to the Greek Testament (Edinburgh 1926/1953) s.v.; cf. TWNT 
s.v. evtoA.r|. 

42 Contra: Turcan 1981a, 118 = 2000, 118. 

13 Cf. Tertullian, Sped. 23.3; Arnobius, Adv. nat. 4.36. C. Panayotakis, Baptism and 
Crucifixion on the Mimic Stage, Mnemosyne 50 (1997) 302—19; R. Lim, The 'Temple 
of Laughter'?: Visual and literary Representations of Spectators at Roman Games, 
in B. Bergmann and C. Kondoleon (eds.), The Art of Ancient Spectacle (New Haven and 
London 1999) 343-65. 



156 CHAPTER THREE 

inner person ridiculed to one's face? And if we should never, of our own 
freewill, exchange our own face (7tp6o"(07tov) for a more ridiculous one 
[i.e. wear a comic mask]. . . . 

Paedag. 2.45.1-3 Marrou-Mondesert 44 

There were, then, implicit norms of behaviour as well as standardised 
ones, though they were not necessarily written down. If this were not 
the case, the gods' role as judges, which I mentioned earlier as part 
of the rites of initiation (in connection with the World Beyond, or the 
selection of candidates by those same gods for initiation), would be 
unintelligible. We have to assume a scale of values applying to an entire 
initiatory community, not simply as a characteristic ethos but as a true 
ethical norm that determined the conduct both of individuals and of 
the group as a whole. In my view, Apuleius' Metamorphoses Bk. 1 1 shows 
us Lucius' progress towards recognition of such a norm. Of course such 
values, even more than myth, are subject to historical change within a 
specific historical formation. The aim of the regulation of social rela- 
tions is to maintain the integrity of the group. Since such regulation 
affects every aspect of social life, some rules will be quite general (what 
some people think of as 'natural prohibitions', such as the prohibition 
of murder or adultery), others will be a function of circumstances. 
These are the group's values, i.e. the behaviour collectively approved 
or disapproved. During the course of their integration into the group, 
the initiands learn how to recognise these, so that the extent of an 
individual's adaptation to the rules is also an index of his socialisation 
into the group. 45 If we accept that there were groups of initiates that 
existed over time and whose members took part corporately in specific 
religious activities, it follows that such groups must have established 
norms of behaviour. As Meeks observes: "The construction of the com- 
munity and of moral norms is a single, dialectical process" (1993, 213). 
It is the community that controls and disciplines its members to make 
sure they adhere to the rules based on the general moral principles, 
and maintain ethical norms. 



" The entire chapter is eloquent, and chilling, testimony to the kill-joy efforts of 
Christian leaders. 

45 On the socialisation of the dominant class at Rome, see for example W.M. 
Bloomer, Schooling in Persona: Imagination and Subordination in Roman Education, 
ClassAnt 16 (1997) 57-78. 



SYSTEMS OF VALUE 157 

Everyday life however produces problems, both as regards the degree 
of acceptance of the values as a whole and in respect of their practical 
application. The passage of time, the number of people involved, the 
authoritative assertion by others of values or behaviour not in accord 
with the internal norms, all these will bring about changes and the 
development of new values. If it is only a matter of one or two, they 
can co-exist with the accepted forms. In fact all moral systems hap- 
pily develop contradictory values; the individual can thus emphasise, 
positively or negatively, the one that fits his immediate needs or that 
may serve him as a general guide to living. Looking at the relevant 
information provided by our documentation might provide interesting 
results in this connection. 

Coming as they did from different parts of the eastern Mediterranean, 
our cults brought a variety of cultural baggage with them, all of course 
heavily conditioned by their integration into the Graeco-Roman context. 
The values implicit in their myths might overlap to some degree with the 
ethical norms dominant in the culture they aimed at settling in; how- 
ever, the specific ways in which appropriate behaviour was inculcated 
were obviously different. This can be seen, for example, just by looking 
at any of the versions of the myths of Isis and Cybele. It is therefore 
impossible, as well as unnecessary, to attempt to draw up a scheme of 
values common to all three. However it might be worth stressing that 
at least sexual abstinence seems to have been a shared value; perhaps 
it was felt that such renunciation made the individual more receptive 
to the divine. At any rate, initiation was preceded by the implicit or 
explicit obligation that one refrain from sexual contact. 

In relation to Mithraism, it is sufficient to recall the gender exclusiv- 
ity of its adherents, to which I have already referred (Chap. 2 n. 290 
above). Surprisingly enough there was no female counterpart to the 
male Mithraist (as the 'Normal Female' to the 'Normal Male'). 46 In 
this context, masculinity was so positively marked that Mithras' obvi- 
ous indifference to women was reproduced in the social world of the 
cult. 47 This however had nothing to do with the rejection of marriage 



46 'Normal Male/Female' are terms used by H.N. Parker, The Teratogenic Grid, in 
Hallett and Skinner 1997, 47-65. 

" Note Bourdieu's point that in what he calls 'rites of institution' what is impor- 
tant is not the passage or transition itself but the nature of the imagined 'line' that 
is crossed. The true boundary is between those in principle eligible for the ritual and 
those who are not (2001, 176). 



158 CHAPTER THREE 

or family-bonds: there are many votives pro se et suis, sons were often 
introduced into the cult by their fathers, and there are at least a handful 
of tomb-stones of Mithraic office-holders put up by their wives (Clauss 
2000, 39). The exclusion of women from the cult was mainly a matter 
of taking Mithras as a model for one's religious life. 

It is relevant here that some Graeco-Roman novels have been inter- 
preted, among things, as coded initiations into marriage and sexual 
relations, 48 taken as metaphors for initiation into the mysteries them- 
selves. 49 In Heliodorus' Aethiopica, for example, the socially-acceptable 
aim in this context, marriage, is presented in terms of an initiation into 
the mysteries. 30 Initiation into the mysteries of Eros is the telos both of 
Longus' Daphnis and Chloe and Achilles Tatius' Leucippe and Clitophon, 
and likewise presented in terms closely modelled on initiation into the 
regular mysteries. 51 It also represents Chloe's exposure to the licentious 



48 It is obvious that marriage was positively marked in Roman ideology, not merely 
from a political perspective but also because the family bond was held both to keep 
society together and ensure its continuity through the blessing of children, cf. Montserrat 
1996, 80fE; S. Dixon, The Sentimental Ideal of the Roman Family, in B. Rawson (ed.), 
Marriage, Divorce and Children in Ancient Rome (Oxford 1991) 99—1 13; eadem, The Roman 
Family (Baltimore 1992); P. Veyne, La famille et l'amour sous l'Haut-Empire Romain, 
Annales ESC 33 (1978) 35-63; S.M. Treggiari, Roman Marriage (Oxford 1991). 

19 I refer of course to the provocative and much-debated thesis of Merkelbach 1962, 
that the novels are a genuinely mystic form of literature, their plots patterned on the 
initiatory experience. Some may be reminded here of the claims about the origin of 
tragedy made by Gilbert Murray, EM. Cornford and Jane Harrison. Analogous claims 
have been made for the influence of Eleusis on Greek lyric poetry: R. Garner, Mules, 
Mysteries and Song in Pindar's Olympian 6, ClassAnt 1 1 (1992) 45—67. Others have picked 
up Merkelbach's basic points but pursued them more indirectly, e.g. M.J. Hidalgo, La 
novela griego como vehiculo de propaganda religioso, in Alvar, Blanquez and Wagner 
1993, 197—214; S. Lalanne, Hellenism and Romanization: A Comparison between the 
Greek Novels and the Tale of Psyche in Apuleius' Metamorphoses, in Ostenfeld 2002, 
225—32; from the point of view of genre: Konstan 1994. In my view, Merkelbach's 
claims are marked by an incomprehensible desire to turn all the novels into romans a 
clef. Allegory is a catching ailment. 

50 Beck 1996, 145. Foucault 1986, 228f. argues likewise that the Greek novel is 
grounded in a 'heterosexual' relation marked by a male-female polarity and an insistence 
on sexual abstention "modelled much more on virginal integrity than on the political 
and virile domination of desires". In his view, the "fulfilment and reward of this purity 
is a union that has the form and value of a spiritual marriage". Such a view is obviously 
very dependent on the Plutarchan view that in marriage to desire is a higher good than 
to be desired: to yap sp&v ev yd|j,cp to\> epaoGai uei^ov aya06v ecmv (Marr. amat. 23, 
769d). According to Foucault (1986, 179-85), the late first century AD saw the rise of 
a different type of erotic ideal from that generally followed in Greek culture prior to 
that time, while retaining certain older constants; Plutarch's aim is to synthesise Plato's 
two types of love into a single Eros. Cf. too Goldhill 1995, 158. 

51 Longus: "As the final sentence reads, Daphnis enacts (eSpaae) what his initiatrix 
Lycaenion had taught (e7rai§e\)ae) him and Chloe learns (euaBev) that what had hap- 



SYSTEMS OF VALUE 159 

world of male sexual violence in terms of initiation. 12 Whatever the 
precise relation between the ancient novel and the mystery-cults, it seems 
clear that in each the motif of a real or imaginary journey or transition 
was used to convey the idea of social integration, in the specific form, 
say, of marriage. The safe accomplishment of this journey, in which 
the ethical qualities of the heroes/protagonists were heavily stressed, 
turned them into ideal ethical models. Such models were there to be 
imitated both by readers and by adherents of our mysteries, disposed 
as they were to become morally irreproachable, not merely by their 
own efforts but also with the aid of divine grace. 53 

It must however be said that every-day life meant that it was quite 
impossible to fulfil all the ethical demands theoretically made by these 
cults (Alvar 1993a). Their adherents suffered the usual vicissitudes that 
are the lot of human-beings. Whatever resolutions they made at the high 
points of mystical devotion, the troubles of everyday life outweighed the 
forces of Utopia. In other words, it was often impossible to attain the 
ideal. 54 Nevertheless I do believe that there was an element of Utopian 
illusion in these cults, deriving from the symbolic 'rebirth' achieved 
through the initiate's symbolic death. Submission to this ritual seems 
to imply acceptance of an Utopian project implicit in the adoption of a 
new set of religious practices, or what might nowadays be called a new 
religious 'identity'. 53 The change of name practised by some initiates 
can be read as a desire to make such a new identity explicit, the moral 
transformation being figured in the new name. 06 This is not the place 



pened in the woods had been mere child's play (itaiyvux)": Beck 1996, 145 n. 61; 
Achilles Tatius: H. Morales, Vision and Narrative in Achilles Tatius' Leucippe and Clitophon 
(Cambridge 2004). 

2 JJ- Winkler, The Education of Chloe: Hidden Injuries of Sex, in idem, The Con- 
straints of Desire: The Anthropology of Sex and Gender in Ancient Greece (New York 1990) 101-26. 
Winkler's argument is rightly criticised by Goldhill as one-sided (1995: 33-45). 
:,:! See my further remarks in Chap. 4.4.b.ii below. 

54 One might compare the Qumran community, which jealously guarded its revealed 
wisdom, accessible only to those who had managed to complete their spiritual trans- 
formation: D. Dimant, The Self-image of the Qumran Community, in A. Berlin (ed.), 
Religion and Politics in the Ancient Near-East (Bethesda 1996) 93—103. On the variety of 
motives for permanent sexual continence in Judaism, note P.W. van der Horst, Der 
Zolibat im Fruhjudentum, in W. Kraus and K.-W. Niebuhr (eds.), Fruhjudentum undNeues 
Testament im Horizont biblischer Theologie. Wiss. Untersuchungen zum Neuen Testament 
162 (Tubingen 2003) 3-14. 

55 For a fine account of the construction of a specifically Christian identity out of 
a predominantly Jewish one, see Lieu 2004. 

56 Cf G.H.R. Horsley Name Changes as an Indication of Religious Conversion in 
Antiquity, Numen 34 (1987) 1-17. One or two cases have been noted in the case of the 



1 60 CHAPTER THREE 

to start a discussion of the religious significance of theophoric names, 
or of pseudonyms, nicknames or agnomina as evidence of religious 
conversion, which requires independent treatment." It is enough here 
to say that such usages may in any given case be evidence of a desire 
to mark becoming a new man, one capable, thanks to initiation, of 
overcoming the 'vices of the world'. 

Becoming a new man seemed necessary because of our mysteries' 
project of establishing a new ethical and moral order, not of course 
for the society as a whole but as a corollary of the truths they asserted. 
But they had in practice to differentiate between the aim and the actual 
changes they could accomplish. Even during the initiatory process the 
individual's hope for radical moral change was conditioned by the real 
possibilities open to him. And afterwards in most cases everyday life 
gnawed away at the experience until nothing was left. Reality inevitably 
imposed itself as the condition under which life had to be lived even 
by initiates. They did not form a group out of this world, cut off from 
links with their society, but were all the while integrated into the life 
of their cities, taking a full part in social life. 

This problem has of course been much discussed from a different 
point of view, namely conversion. 58 The discussion has revolved around 
the question of whether initiates into the mysteries experienced a form 
of conversion. As so often, there is not much evidence from antiquity 
on this point. The positions taken up by scholars thus tend to reflect 



cult of Isis, e.g. Witt 1971, 307 n. 39; 311 n. 16; 321 n. 55; Malaise 1972a, 29. Of all 
changes of names to mark a change of status, the most famous for a Spaniard is the 
re-naming of Don Quixote's horse, Rosinante, though unfortunately its result was the 
opposite of what was intended: 'And so, after having composed, struck out, rejected, 
added to, unmade, and remade a multitude of names out of his memory and fancy, 
he decided upon calling him Rocinante, a name, to his thinking, lofty, sonorous, and 
significant of his condition as a hack before he became what he now was, the first and 
foremost of all the hacks in the world" (tr. J. Ormsby). 

57 See the brief but important remarks of R.C.T. Parker, Theophoric Names and 
the History of Greek Religion, in E. Mathews and S. Hornblower (eds.), Greek Personal 
Names: Their Value as Evidence (Oxford 2000) 53-79. 

58 Nock, 1933 flatly denied the possibility of true conversion outwith Christianity, 
since it was only in that context that the convert was required to reject other deities, 
and acceptance of the postulate of a single divinity became a decisive criterion of the 
authenticity of a change of heart. More recent discussions from the point of view of 
the sociology of religion include: B. Kilborn and J. Richardson, Paradigm Conflict: 
Types of Conversion and Conversion Theories, Sociological Analysis 50 (1988) 1—21; 
C. Ullman, The Transformed Self: The Psychology of Religious Conversion (New York 1989); 
L. Rambo, Understanding Religious Conversion (New Haven 1993); A. Oksanen, Religious 
Conversion. A Meta-Analytical Study (Lund 1994). 



SYSTEMS OF VALUE 161 

their a prioristic assumptions. Since the best-known case is that of the 
fictional Lucius in Apuleius' Metamorphoses, he has tended to be the 
focus of interest.' 9 His change of view with regard to the order of 
the cosmos and the powers that regulate it, and to the moral claims 
made upon him, quite apart from the change in his ritual life (which 
is not at issue here), are sufficiently explicit for us to conclude that he 
did experience a transformation or change of heart, which we have no 
trouble in locating within the semantic range of our word conversion. 
It is a different matter if we insist on a strict definition of the word 
in the direction of exclusivity, which is really only appropriate to the 
hegemonic claims of Christianity, eager as it was to avoid any contagion 
from an earlier mode of religiosity. The case of Lucius, which I take 
to be typical of many other initiates, did not involve the abandonment 
of other religious practices: initiation was understood as a means of 
deepening earlier religious experiences. We might therefore consider it 
to be not so much a conversion experience as a mystical option within 
the religious system, a specialised version of henotheism (Bradley 1998, 
331). There is no simple solution here, since a great deal depends on 
what one considers conversion to involve. If we take the 'mystical' 
tack, it surely is the case that what we may characterise as henotheistic 
tendencies were something new in Roman religious practice, involving 
a new way of conceiving divinity, which we might call 'henotheistic 
conversion'. Henotheism did not involve rejection of the other gods 
but claimed unequivocally that the god in question was superior to all 
others. Such a claim surely amounts to a true conversion. 

However that may be, in my view the oriental cults, with their mys- 
tery component, became so integrated into the society of the Empire 
that from a fairly early point the members were not recruited through 
conversion or selection but through family membership: they were 
largely the children of the families where this type of religiosity was 



19 For the debate since Nock, I refer to the careful discussion of Shumate 1996. 
Note also the replies to R. MacMullen, Christianizing the Roman Empire AD 100—400 (New 
Haven 1984) and idem, Conversion: A Historian's View, The Second Century 5 (1985—86) 
67-81, by W. Babcock, MacMullen on Conversion: A Response, ibid. 82-89; M.Jor- 
dan, Philosophic 'Conversion' and Christian Conversion: A Gloss on Prof. MacMul- 
len, ibid. 90—96. Other recent treatments: A. Malherbe, Paul and the Thessalonians: The 
Philosophic Tradition of Pastoral Care (Philadelphia 1987) 21-28; E. Gallagher, Expectation 
and Experience: Explaining Religious Conversion (Atlanta GA 1990) 109-33; M. Goodman, 
Mission and Conversion: Proselytizing in the Religious History of the Roman Empire (Oxford 
1994). T.M. Finn, From Death to Rebirth: Ritual and Conversion in Antiquity (Mahwah NJ 
1997) can cheerfully be ignored. 



162 CHAPTER THREE 

favoured. 60 If that were the case, the significance of the issue of conver- 
sion would be greatly reduced. There is another point too in the same 
connection. Although these cults clearly did have their own accounts 
of the real world (which the convert would tend to prefer), their degree 
of integration with the wider culture makes it very difficult for us to 
differentiate meaningfully between the various systems. 61 To return to 
the issue of conversion, it may be hard to define the term strictly, but 
it certainly ought not to be confused with religious exclusivism, which 
is characteristic only of one type of conversion, a type that there is 
no reason to take as more authentic than others, or more universally 
valid. The idea that the goodness of one deity far exceeds that of oth- 
ers, and the worshipper's preference for that deity, can perfectly well 
be interpreted, for example in the case of Lucius, as amounting to a 
henotheistic conversion. Where we do find substitutions, they are partly 
based on analogy — the antagonism of neighbours — and partly competi- 
tive: henotheism feeds on ambiguities of this kind, which make it very 
difficult to interpret such cases correctly. A case in point is the re-cut 
inscription on the front of a very small marble cippus found in 1912 in 
the pronaos of the mithraeum beneath the outer substructures of the 
Antonine Baths, where the name of Serapis has been replaced by that 
of Mithras: eiq Zzvq I Mvxpcxc; I "HAaoc; I KoaixoKpaxcop I dveiKT|i:o<;. 62 
The case is made more complicated by the fact that there is another 
inscription on the reverse, this time a regular invocation of Zeus Helios 



60 Cf. R. Wiegels, Die Rezeption orientalischer Kulte in Rom — UmriB eines 
Forschungsfeldes zum Thema: Religion und Gesellschaft in romischer Zeit, Freiburger 
Universitatsblatter 65 (1979) 37-61; J. Alvar et al., La religiosidad misterica en el espacio 
familiar, ARTS 1 (1998) 213-25. 

61 The final section of Nancy Shumate's book argues that the world Lucius ends up 
by joining is culturally quite different from that of Rome. Well though I think of the 
remainder of her book, such a conclusion is surely quite misguided. 

62 AE 1913: 118 = V. 463 = IGVR 194a = RICIS 501/0126 (first quarter IIP). 
According to Ghislanzoni 1912, 323, the cutter first wrote MPAS (this and the T for 
perhaps suggests some unfamiliarity with the name). The fact that the re-cutting was 
actually over the name Serapis was first noticed by L. Canet ap. Cumont 1919, 319; 
cf. Cumont 1929, 79 fig. 5. This face is taken to be the front of the cippus because the 
moulding is on that side. The stone is very small and easily transportable; in my view 
it originally came from elsewhere before being re-used. Many of the hypotheses that 
have been suggested in this connection seem to me quite implausible, based as they 
are on unargued, and arbitrary, assumptions: Nock for example suggested (1925, 89) 
that it provides evidence of a sort of damnatio memoriae in connection with the murder 
of the philo-Egyptian Caracalla; following Cumont 1929, n. 37, Simon 1978, 468 
suggested that it is to be understood as evidence of a certain rivalry between the cults 
of Sarapis and Mithras. I am sceptical (see n. 66 below). 



SYSTEMS OF VALUE 163 

Sarapis, where the name of Mithras has been inserted near the end, 
evidently prompted by the epithet aveticriToc;. 63 The cemented pit in 
the centre of the floor of this mithraeum (1.90m deep), and the subter- 
ranean connecting passage to the adjoining room, which are unique in 
the cult of Mithras, might indicate that this part of the substructures of 
the Baths had in fact previously been used by another, wealthy cult. 64 
In my view, the likeliest explanation of the cippus in its present form is 
that an individual made two different attempts, at a time (say the second 
third of the third century AD) when Mithras was widely understood 
as a demiurge and kosmokrator on a par with Zeus, to convey this 
understanding in terms already established for solar Serapis. 65 This 
may also be the best explanation for the presence of busts of Serapis 



63 IGVR 194b: Au 'HiUcp ixeydXcp EapdjuSi aartfipi ;iXoi)to86ti (sic) cttikog) euepyeTri 
dveiKriTtp MiGpa xapiairipiov. There is now a sharp division over the interpretation 
of this text between those who work on the Egyptian cults and those who work on 
Mithras: Malaise 1972a, 465; Vidman 1970, 148; RICIS 501/0126 all claim that MI0PA 
here is the name of the dedicator, evidently seduced by the name of the Isiac priest 
in Apuleius, Metam. 11.13, 15 etc., which they always cite. However this is a desperate 
remedy, because, epigraphically the personal name seems always to be Mithres, not 
Mithras (for example, of the 19 cases of the name listed by M. Bang in the name- 
index to CIL VI, irrespective of his absurd system of brackets and doubts, not one 
ends in -as, all in -es, including the case Bricault cites to support his argument: CIL 
VI 571 = RICIS 501/0125) and it assumes a complete divorce between the reverse 
and the face. Simon 1979, 414 thought, implausibly enough, that the effect of placing 
Mithras in this position was to give him particular emphasis. There can surely be no 
question that in fact the trio Zeus-Helios-Serapis is primary, with Mithras being added 
as an afterthought. 

64 A.D. Nock 1925, 88 n. 35; 1937, 113, in keeping with his odd belief that there 
was a close connection between the Phrygian cults and those of Mithras, thought that 
this structure was proof that the taurobolium was also practised in the cult of Mithras. 
D.M. Cosi, II mitreo nelle Terme di Caracalla: Riflessioni sulla presunta fossa sanguinis 
del mitreo delle Terme di Caracalla, in Bianchi 1979, 933-42 (with excellent photos, 
pis. XXVI-XXX) showed how implausible this was, and suggested that the hole served 
as a place from which ritual 'events' could be presented. After re-examining the site, 
M. Piranomonte has recently revived the idea that the pit was for the taurobolium 
(ap. LTUR 3 [1996] 267f. s.v. Mithra: Thermae Antoninianae, and further lit.), but 
her argument is completely circular: as we shall see, since McLynn's fundamental 
article (McLynn 1996) it is untenable to claim that the taurobolium involved the 
ritual described, or rather invented, by Prudentius (see Chap. 4.3c), so that some other 
explanation must be preferred. Very little was found in the mithraeum, and we have no 
idea when it may have been turned into one. Alternatively, the subterranean chamber 
may have been used as a temporary storage for the rubbish from meals; the narrow 
corridor that led down to it started between two latrines. 

65 Cf. Simon 1979, 413f.; Gordon 2006: 189-93. The first version of the text is 
usually dated to the reign of Caracalla 'or a little earlier', but in my view there is no 
good reason for assigning such a date. Henotheistic tendencies in Mithraism have been 
studied by M. Clauss, Omnipotens Mithras, Epigraphica 50 (1988) 151-61. 



1 64 CHAPTER THREE 

in mithraea (though all the cases reported so far are in different ways 
problematic). 66 

As I pointed out earlier, the process of integration was gradual. Ini- 
tially, in the Late Republic, the new religious practices, at any rate the 
cult of Isis, were rejected by Romans of the elite. Although this hostility 
was provoked by essentially political anxieties about the instrumentali- 
sation of such new cults by enemies of the ruling order, the successive 
expulsions on the orders of the Senate suggest that they were felt to 
be deeply alien. As the plebs was tamed, however, and emperor-wor- 
ship began to be seen as an effective means of re-ordering the religious 
system as a whole, such criticism lost its force. It is generally supposed 
that Claudius reformed the Phrygian cult (though I am sceptical), and 
both Caius and the Flavians moved to assimilate the Egyptian cults. 67 
We may assume that throughout this period the cults themselves made 
an effort, conscious or not, to rid themselves of their alien or marginal 
features and emphasise their integrative ones. This amounted to an 
unconscious adaptive strategy, furthered by the fact that they suffered 
neither internal theological disputes nor secessions due to purist zeal. 
Their ethical rules, and their acceptance of social difference, threatened 
no one. Eventually they proved useful in offering resistance to Christian- 
ity, and the official objections ceased entirely. 

There was however a counter-view, that not only failed to discern 
any moral superiority on the part of the adherents of these cults, but 
saw them as the fount of corruption of basic Roman institutions, such 
as the family. The tradition of satire dealt precisely in such violently- 
drawn contrasts between an idealised 'us' and a corrupted 'them', with 
women often in the middle, serving as a fifth column. A good example 
is Juvenal's sixth satire, a misogynistic diatribe against marriage, where 
women are represented as capable of any subterfuge and folly in order 



66 Compare M. d'Asdia, Mosaici nilotici dal Celio, Bollettino d'Arte 109-10 (1999) 
77—86 at 82 n. 28. Although there are no other inscriptions that mention both Serapis 
and Mithras (V 792 is irrelevant, since it was part of a cache of concealed statues at 
Emerita, by no means all of them Mithraic), busts of Serapis have been found in two 
mithraea, S. Prisca in Rome (V 479) and the Walbrook in London (V 818). In both 
cases caution is called for, since the bust at S. Prisca might belong to the enormous 
quantities of infill, and the group at the Walbrook has other peculiarities (cf. R.L. 
Gordon, JRA 13 [2000] 736-42). The case(s) at Emerita (V 783, 787) are also very 
doubtful, for the reasons given above, likewise the head of Isis found in the mithraeum 
of the Castra peregrinorum on the Caelian (Lissi Caronna 1986, 39). 

67 Cf. Fishwick 1966; Vermaseren 1977, 122f; Takacs 1995. 



SYSTEMS OF VALUE 1 65 

to make their husbands' lives hell. 68 Among these is joining one of 
those dreadful new cults that attract people of low intelligence, and 
are good only to ruin a family's fortune The wife in this scenario is 
simply a credulous dupe: 

Look! In comes the troupe of frenzied Bellona and the Mother of the 
Gods, along with an enormous eunuch, a face his perverted sidekick must 
revere. A long time ago now he picked up a shard and cut off his soft 
genitals. The noisy band and the common drums fall quiet in his presence 
and his cheeks are clothed in the Phrygian cap. In a booming voice he tells 
the woman to beware the arrival of September and the southerly winds, 
unless she purifies herself with a hundred eggs and presents him with 
her old russet-coloured dresses, to ensure that any serious or unforeseen 
disaster that's impending disappears into the clothes and atones for the 
whole year in one go. In the winter-time she'll break the ice, step down 
into the river and submerge herself three times in the morning Tiber, 
even cleansing her head in those swirling waters. Then, naked and shiv- 
ering, she'll crawl right across the Proud King's Field on bleeding knees. 
If white Io tells her to, she'll go to the ends of Egypt and bring back 
water fetched from sweltering Meroe to sprinkle in Isis' temple, towering 
next to the ancient sheepfold [the Saepta in the Campus Martius]. You 
see, she think her instructions come from the voice of her Lady herself! 
There you have the kind of mind and soul that the gods converse with 
at night! Consequently the highest, most exceptional honour is awarded 
to Anubis, who runs along mocking the wailing populace, surrounded 
by his creatures in linen garments and with shaved heads. He's the one 
that asks for a pardon whenever your wife does not refrain from sex on 
the days which should be kept sacred and a large fine is due for violation 
of the guilt. When the silver snake has been seen to move its head, it's 
his tears and his practised mumblings which will ensure that Osiris will 
not refuses to pardon her fault — provided, of course, he's bribed by a fat 
goose and a slice of sacrificial cake. 

Juvenal, Sat. 6, 512—41, tr. S. Braund 



2. Ethics in the Phrygian Cults 

II n'est aucune religion orientale dont nous puissions suivre a Rome 
l'evolution progressive aussi exactement que celle du culte de Cybele 



68 "The warning against marriage should... not be read as a recommendation not 
to get married because of the dangers for the individual of becoming a hen-pecked or 
cuckolded husband, but as a general statement of the decay of society: it is useless to 
try and live according to old rules and traditions since their content has changed for 
the worse, and they no longer constitute the bulwark of society": G. Viden, Women in 
Roman Literature: Attitudes of Authors under the Early Empire (Vastervik 1993) 160. 



1 66 CHAPTER THREE 

et Attis, aucune ou apparaisse aussi nettement Tune des causes qui ont 
amene leur decadence commune et leur disparition. Toutes remontent 
jusqu'a une epoque lointaine de barbarie, et elles ont herite de ce passe 
sauvage une foule de mythes dont l'odieux pouvait etre dissimule, mais 
non supprime, par un symbolisme philosophique, de pratiques dont toutes 
les interpretations mystiques deguisaient mal la grossierete fondamentale, 
survivance d'une rude naturalisme. Nulle part la discordance entre les ten- 
dances moralisantes des theologiens et l'impudicite cruelle de la tradition 
n'est aussi eclatante. Un dieu dont on pretend faire le maitre auguste de 
l'univers etait le heros pitoyable et abject d'une obscene aventure d'amour; 
le taurobole, qui cherche a satisfaire les aspirations les plus elevees de 
l'homme vers la purification spirituelle et l'immortalite, apparait comme 
une douche du sang qui fait songer a quelque orgie des cannibales. Les 
lettres et les senateurs qui participaient a ces mysteres y voyaient officier 
des eunuques maquilles, a qui on reprochait des moeurs infames et qui 
se livraient a des danses etourdissantes rappelant les exercises des dervi- 
ches tourneurs et des A'issouas. On comprend la repulsion qu'inspirerent 
ces ceremonies a tous ceux dont le jugement n'etait pas oblitere par 
une devotion fanatique. II n'est aucune superstition de l'idolatrie dont 
les polemiste chretiens parlent avec un mepris plus outrageux, et sans 
doute avec raison. Mais ils n'etaient plus contraints, eux, de verser leur 
vin nouveau dans de vieilles outres, et toutes les ignominies qui purent 
entacher cette antique religion phrygienne ne doivent pas nous rendre 
injustes envers elle et nous faire meconnaitre les longs efforts tentes pour 
l'epurer peu a peu, pour lui donner une forme qui lui permit de repondre 
aux exigences nouvelles de la morale, de suivre la marche penible de la 
societe romaine vers le progres religieux. 

This is how, in the final paragraph of his chapter on the Phrygian cults 
in Les religions orientates, Franz Cumont perceived the system of values 
purveyed by the cult (Cumont 1929, 67f.). He was in his day by no 
means alone; such views were in fact commonplace at the time. The 
technique was to smother the mythical or historical past in a custard 
of ethical values dominant in their own social circle. Stressing Phry- 
gian barbarism and its assault upon civilised, that is to say Hellenised, 
Rome was tacitly to invoke the behavioural norms of a particular social 
group and its interpretation of social reality (Cumont was the scion of 
a wealthy family of Belgian industrialists, a rentier with a private income 
that enabled him to write — lucky man — without being employed by 
an institution). They thus confirmed their own values by more or less 
unconsciously manipulating a past where they discerned a reflection of 
themselves. Convinced of the objectivity of their version of this past, 
they elevated their tautologies to the status of scientific conclusions, 
which then formed the basis of an idealist anthropology. At the same 



SYSTEMS OF VALUE 167 

time, field-work was supposed to provide systematic confirmation of the 
universal validity of these theories by overcoming their temporal and 
spatial limitations. Religion and thought could thus be abstracted from 
actual cultural systems and their 'true nature' discerned. The masters 
of such analytic truths were thus empowered to pass judgement on 
the ideas and praxis of the various 'others', without an inkling of the 
subjective nature of what they were doing. 

Denouncing the moral turpitude of the oriental cults, especially Mater 
Magna, can thus been seen as a strategy for ignoring counter-evidence 
concerning their actual values. Such evidence could be found in the 
ancient texts and was in fact well-known to the cult's modern detractors 
(cf. Sfameni Gasparro 1981, 404). Thus the emperor Julian writes, in 
a manner even they could hardly find inacceptable: 

r\\uv ovv 6 9ea|i6<; napaKeXevexai, xoiq <pvaei |iev oupavioiq, ei<; yfjv 
8e evejcOeiaiv, apexriv |ieta evae^eiac, arco tr\c, ev tf\ yfi noXixeiaq 
d|ir|aa|ievo\)(; 7tapa tf|v 7tpoyoviKTiv kou ^cooyovov anex>8ew 0e6v. 69 

Or. 5, 169b 

A few lines earlier, he also observes that the ceremony of the felling 
of the pine at the beginning of the festival arbor intrat, which we shall 
look at in more detail in the section on Metroac ritual (Chap. 4.3d), 
symbolises the harvesting of the most beautiful thing we have to offer 
the deity, namely virtue and piety (apexf|V nexoc zvatfiziaq). They are 
to be offered to the goddess, oi)|i(3oA,ov xfj<; evxod)9cx xpriaxfjc; noXixziaq 
eaoixevov, as a symbol of our well-ordered polity here on earth (Or. 
5, 169a). Nor is the evidence for such positive values confined to the 
late-antique period. There are also traces of them at a period when 
the Phrygian cults do not even seem to have had a mystery component. 
Lucretius provides a long description of the cult in his day, which can 
be analysed in a variety of different ways (2.600-60). What interests 
me here is the stress laid on filial duty and love (604f, 614—17, 643). 70 



69 "Accordingly the ritual enjoins us, who by nature belong to the heavens but have 
fallen to earth, to reap the harvest of our constitution here on earth, namely, virtue 
and piety, and then strive upwards to the goddess of our fore-fathers, to her who is 
the principle of all life" (tr. W.C. Wright). 

70 See the clear discussion by R.W. Sharpies, Cybele and Loyalty to Parents, Liverpool 
Classical Monthly 10.9 (November 1985) 133f. He also mentions two problematic passages, 
Aristotle, De mirab. ausc. 162, 648b3 and [Plutarch], De fluv. 9.5. The latter describes 
how pious sons with their fathers would collect beryls on Mt Sipylos and place them 
in the temple of the Mother of the Gods, so as to maintain faith with their ancestors. 



1 68 CHAPTER THREE 

The passage also stresses the fear instilled in the vulgar crowd by the 
sight of the goddess in procession (622f)," and points to the role of 
self-discipline, piety and loyalty to one's country as moral virtues char- 
acteristic of the cult. 72 To stress exclusively the shocking aspects of the 
cult is thus extremely one-sided. 

I think the most promising method in this case is to take another 
look at the myth from this point of view, reading it as a means of com- 
municating in symbolic terms a specific notion of transgression. Julian 
himself seems to do this at Or. 5, 169cd, where he interprets the myth 
of Attis as an allegory of the creation of the world, and the rituals 
as an invitation to piety, so that the souls of the worshippers may be 
restored to their celestial home, whence they came, and where they are 
to be re-united with the One. 

One issue has to do with the sexual regime of the family, another 
with the difference between the divine world and the human, a third 
with dropping out. These at any rate can be seen as three types of 
transgression that are criticised and sanctioned in the myth in the name 
of preserving the established social order. 73 

First, the family. The function of the ban on mother-son incest is 
to protect the structure of the patriarchal family, based as it is on the 
notion of exclusive ownership of the female. 74 For a man to undermine 



The appearance of the same theme in Lucretius seems to indicate that there may be a 
grain of truth in this, despite the well-known problems posed by Defluv. For Mt. Sipylos 
near Smyrna as a centre of the worship of Cybele, cf. Strabo, Geogr. 10.3.12, 469C; 
Pausanias 3.22.4, 5.13.7; CCCA 1 nos. 544-47, 550, 555, 564, 571, 583; F. Biirchner, 
s.v. Sipylos 1, RE 3A (1927) 275-81; Roller 1999, 200-02. 

71 Ingratos animos atque impia pectora volgi conterrere metu quae possint numine divae, "that they 
may amaze the ungrateful minds and impious hearts of the vulgar with fear through 
the goddess' majesty" (tr. M. Ferguson Smith). 

' 2 Cf. Summers 1996, 341. I do not however understand why he assumes that it must 
have been the Roman nobility who "promoted the moral lessons being drawn from it 
[i.e. the cult of Cybele]", whatever quite we are to understand by that. 

73 Cf. Cosi 1986. Three of the contributions to AA.VV 1981a are particularly rel- 
evant here: J. Scheid, Le delit religieux dans la Rome tardo-republicaine, pp. 1 1 7-7 1 ; 
D. Sabbatucci, II peccato 'cosmico', pp. 173-77; T Cornell, Some Observations on 
the crimen incesti, pp. 27—37. On the latter, see also M. Beard, Re-reading (Vestal) Vir- 
ginity, in R. Hawley and B. Levick (eds.), Women in Antiquity: New Assessments (London 
1995) 166-77. 

74 It is of course not my intention to impose a universal category (since, as Levi- 
Strauss showed in Les structures elementaires de la parente [Paris 1949], in view of its uni- 
versality in relation to the theme of exchange, incest may well precede the patriarchal 
family; cf. R. Fox, The Red Lamp of Incest [London 1980] ), simply one that is of use 
in discussing the societies in which the myth of Attis was picked up. There may be 
demographic and genetic reasons for avoiding incest (e.g. R. Fox, Kinship and Marriage: 



SYSTEMS OF VALUE 1 69 

the accepted structure of relations within the family in such a manner 
is to deserve to lose his genitals, the equivalent of being condemned to 
death (as at Rome: Cicero, Leg. 2.23). 75 However there is also another 
sort of danger, represented concretely by Agdistis bursting in upon the 
marriage-feast. I take this to refer to adultery, the groom being hardly 
in a position to refuse the offer. The divine nature of the seductress 
means she cannot be punished, even though the death of the desired 
object causes her intense grief and regret. It is however also thinkable 
that Agdistis' earlier loss of her hermaphroditism was in fact itself 
a violent punishment, which then triggered off the ensuing tragedy. 
However that may be, the message is clear: the brazen attempt at 
seduction produces a train of disaster for the bridal pair, who cannot 
even consummate their marriage. 76 

Second, the difference between the divine world and the human. Attis 
as herdsman succumbs to the advances of a goddess. Here the point is 
not so much their blood-relationship (mother-son) as the transgression 
of the divine-human boundary: Attis' conceit in breaking this rule and 
elevating himself to divine status is punished. At this level, the hidden 
agenda is to reinforce the fixed quality of the order established in society 
and thus maintain the status quo. In the myth however the sanctions 
are imposed asymmetrically: only the one who attempts to rise in status 
is punished. In other words, the mighty, like the gods, their reflection, 
can transgress social norms with impunity: the myth has them freely 
entering into sexual relations that transgress both the status-boundary 
and the marriage-bond. On the other and, I do not think the myth is at 
all interested in masturbation, as Loisy believed (1930, 97) — he was after 
all old enough to have been influenced by the late nineteenth-century 



An Anthropological Perspective [Harmondsworth 1967]; N.B.M. Mead, s.v. Incest, in IESS 
7 [1968] 1 15-22), but they are largely irrelevant here. 

75 A myth current among the Tatuyo of Columbia is helpful in conceptualising 
the problem here: P. Bidou, On Incest and Death: A Myth of the Tatuyo Indians of 
Northwest Amazonia, in M. Izard and P. Smith (eds.), Between Belief and Transgression: 
Structuralist Essays in Religion, History and Myth (Chicago 1982/Paris 1979) 129-51. 

76 Cf. G. Hoffmann, Le chdtiment des amants dans la Grece classique (Paris 1990) 52—77. 
My point about the ambiguity of Agdistis seems in tune with his remarks on p. 68: 
"Loin de s'affirmer comme un seducteur doue d'une virilite au-dessus de tout soupcon 
et de vouloir a tout prix s'opposer au femmes qui leur sont attributes, l'adultere grec, de 
Paris a Alcibiade, joue d'une certaine ambiguite. Archer des lisieres, adulte immature, 
il est aussi un etre equivoque, comme s'il visait a reunir en sa personne les attraits des 
deux sexes et a exasperer chez ses admirateurs le desir toujours insatisfait de la fusion 
en un de deux corps incomplets". 



170 CHAPTER THREE 

European anxieties about male masturbation — given the fact that the 
emission of sperm occurs while Zeus is dreaming. 

The third theme is dropping out, which is seen as a punishable 
betrayal. I understand this in a religious sense. The human-being, 
chosen by the god, enjoys a hierogamy or at any rate agrees to enter 
her service, despite the fact that this restricts his freedom of action and 
of movement. He enters a sort of metaphysical temenos that diminishes 
his right to social intercourse (cf Julian, Or. 5, 167a). He can no longer 
return with impunity to the social world from which he has emancipated 
himself, as though his privileged status allowed him to ignore the divine 
will. In this context, it is interesting to note how, in Catullus' poem 
(63), Attis tries to escape the goddess' control despite having already 
cut off his own testicles (cf. Nauta 2004). He regrets what he has done: 
iam iam dolet quod egi, iam iamque paenitet (73). He wants to return to a 
past he is now cut off from, but Cybele prevents him from doing so by 
setting one of her mighty lions on him, which makes sure he cannot 
quit the sacred grove: 

fac utifuroris ictu reditum in nemora ferat, 

mea libere nimis quifugere imperia cupit (79f.). 77 

Catullus' scenario highlights the transgressive character of Attis' longing 
to escape his service. 78 This longing to return to the familiar world is 



77 G.N. Sandy, The Imagery of Catullus 63, TAPhA 99 (1968) 389-99; P. Fedeli, 
Attis e il leone: dall'epigramma ellenistico al c.63 di Catullo, in Letterature comparate. 
Problemi e metodo. Studi in onore di E. Paratore (Bologna 1981) 1: 247—56; Takacs 1996, 38 If. 
J. Godwin, Catullus. Poems 61-68 (Warminster 1995) 12 If, denies both that the poem 
has any moral intention and that it has any implications for marriage. The real difficulty 
is that there is no 'logic' to the myth of the kind that such commentators imagine: 
they are hunting for an aetiology, an answer to a question they cannot even formulate, 
namely: why do the priests of the Phrygian cults castrate themselves? 

78 During the period when rites of passage were all the rage, several scholars sug- 
gested that the point in Catullus 63 is rather that Attis refuses to, or cannot, pass from 
youth to adulthood. O. Thomsen, Ritual and Desire. Catullus 61 and 62 and Other Ancient 
Documents on Wedding and Marriage (Aarhus 1992) 98 argued that Attis is escaping from 
sexual maturity, i.e. refusing to become a. paterfamilias; he discerns a 'pederastic melan- 
choly' within adult sexuality. This is direcdy contradicted by Attis' regret for the civic 
life, the social status, and the property (bona) he has had to abandon (59-67), cf. Takacs 
1996, 381. It has also been suggested that Attis is unable to transcend his youthful 
status in order to achieve the heterosexuality appropriate to adulthood: J. Strauss Clay 
1995. Again one can object that Attis specifically regrets the children he can never have 
(69: ego sir sterilis ero?). An interesting interpretation of Attis as a liminal figure thereby 
capable of integrating opposites is to be found in B.M. Naasstrom, The Abhorrence of 
Love. Studies in Rituals and Mystic Aspects in Catullus' Poem of Attis (Uppsala 1989). 



SYSTEMS OF VALUE 171 

represented as occurring immediately after the self-castration. Catullus 
thus imaginatively enters into the mind of the extreme adherent, sug- 
gests the conflicts experienced by those induced to perform this act upon 
themselves: once they have done it, they must feel regret and, their eyes 
bathed in tears, long vainly to put the clock back, just as their god was 
prevented from returning to his own home, to his place of origin: 

liquidaque mente vidit sine quis ubiqueforet, 
animo aestuante rursum reditum ad vada tetulit. 
ibi maria vasta visens lacrimantibus oculis, 
patriam allocuta maestast ita voce miseriter (46— 49).' 9 

In the myth, as opposed to Catullus' poem, the motif of escape is 
expressed through the idea of wanting to marry a mortal woman: such 
an act would return Attis to normal life. This desire too is represented as 
a wrong-doing that has to be punished. Thus Ovid makes Attis exclaim 
that only the shedding of blood can atone for his guilt: 

Merui! Meritas do sanguine poenas. 
Ah, pereant partes quae nocuere mihi! 

Fasti 4.239f. 

I think this message was understood by the more dedicated worshippers, 
those whose commitment was sufficient to make them long for their 
own moral regeneration, the inner counterpart of the physical altera- 
tion they had imposed on their body (the bans on self-castration by 
Roman citizens imply of course that people of this status did perform 
this act on themselves sufficiently often to attract imperial disapproval. 
The ban seems anyway not to have applied to peregrines). 

It thus seems reasonable to read the myth (and even Catullus' poem) 
as an ideal account of priestly self-dedication to the divine world and 
of the punishment that attends the violation of that ideal. The elect 
individual is removed from the realm of ordinary relationships within 
the community so that he may take advantage of the divine offer of life. 
By marrying (or longing for his old life), Attis is seeking to escape from 
his dependence on the goddess, and she inflicts a fearful punishment 
upon everyone concerned. 80 Expressed in different terms, he is casting 
doubt on the primacy of the divine order, casting off religion for the 



79 The gender of Attis, and the gallae/galli, is deliberately unstable throughout the 
poem, cf. Takacs 1996, 377£; Nauta 2004, 601-05. 

80 In Ovid's version there was no prior sexual relation between Cybele and Attis: 
turrigeram casta vinxit amore deam (224). The fault lies plainly with Attis: fallit et in nympha 



172 CHAPTER THREE 

world. This is tantamount to rejecting the ideological superstructure 
whose ultimate function is to paper over the tensions arising from the 
contradictions of the system. 

But there can be no virus-free areas here: the conflict clearly threatens 
the established order of things. The exemplary punishment, the violence 
and coercion inflicted by the dominant group (gods or humans) in order 
to maintain their privileges, is not incompatible with a touch of pity: 
the wretched sinner, having returned to the fold of order, is allowed 
eternally to magnify the greatness of his masters in that his hair never 
stops growing and his little finger can wiggle (as the sources say), or by 
unfeignedly loving the goddess who has removed from him the pleasures 
of sex. Everyone compensates for his misery as best he can. 

To summarise, then, I argue that, alongside interpretations that focus 
on the system of beliefs (imposing order on chaos) or on ritual (as a 
model for praxis), the myth may be analysed as a means of getting 
access to certain values. We may take these as characteristic of the 
ethical systems current in the historical formations that took over the 
story of Attis, the story of death from divine love. One of these values 
is that adult sexual renunciation is morally superior to conjugal life. 81 
The negation of male sexuality is thus parallel to, even a figure for, 
the rejection of geographical rootedness, of work, of the acquisition 
of wealth, and of membership in a specific social group capable of 
reproducing itself. The rejection of ordinary production and repro- 
duction creates a new sort of hierarchy, whose basis is not status as 
the world sees it, and certainly not the trappings of status as ideally 
represented in the corporeal language of antiquity, but the extent of 
one's capacity and preparedness to imitate the god, and, by so doing, 
replicate his devotion to the Great Goddess beyond him. Ideologically, 
then, self-castration is absolutely central to the Phrygian cults, a central- 
ity that surely explains the importance from the second century of the 
institution of the tauro- / criobolium. But the act of castration itself, like 
the self-laceration, is exemplary; the self-sacrifice of the few guarantees 
the authenticity of the commitment of the many 82 



desinit esse quodfuit (229£). Here, Cybele's concern is with the contamination 
represented by sexual contact with women. 

81 Cf. Attis' call to the gallae: agite ite . . . (qui) corpus evirastis Veneris nimio odio, hilarate erae 
citatis erroribus animum (i.e. of Cybele): Catullus 63.12-18. 

82 Three of the new defixiones from the temple-area of the Magna Mater in Mainz 
use imagery from the cult; each refers to the act of self-castration (praecidere; abscidere), 



SYSTEMS OF VALUE 173 

I have also tried to bring out the hope that people in daily life might 
come to recognise the greatness of the Phrygian goddess. For there is 
no dearth of passages in which she is damned as immoral, a charge 
that is in fact, as we shall see, far more applicable to Isis. Predictably 
enough, it is Juvenal, his pen hard and sharp as ever, who takes it upon 
himself to offer such an inverse image of the Mother: 

Hie nullus verbis pudor aut reverentia mensae, 
Hie fturpisf Cybeles etfracta voce loquenti 
libertas et crine senex fanaticus albo 
sacrorum antistes, rarum ac memorabili magni 
gutturis exemplum conducendusque magister. 
Quid tamen expectant, Phrygio quos tempus erat iam 
more supervacuam cultris abrumpere carnem? 83 

Sat. 2, 110-16 

Isis was thus by no means the sole target of the lubricious imagina- 
tion of writers hostile to the introduction of foreign cults. In this case, 
the goddess' followers are made out to be interested solely in having 
a good time in the flesh-pots: their bad manners and vulgar language 
betray their obscure social origins (cf. Pachis 1996). 84 This reputation 
for gluttony may be due to a custom that, looked at in another light, 
might equally be thought of as praiseworthy. During the spring festival 
of Cybele (April 4th- 10th), feasts were arranged, to which guests were 
formally invited. This custom started out as an aristocratic, indeed 
patrician, habit, for which the name was mutitationes: at this season they 
would invite one another to their houses in turn, in commemoration 
of the arrival of the goddess in Rome (Ovid, Fasti 4.353-56; Aulus 
Gellius, JVA 2.24.2). This was of course a means of strengthening 



but are at least as interested in the frenzied self-laceration of the galli and the blood 
that flowed from such behaviour; cf. Blansdorf 2008, nos. 16—18. 

83 "Here there is no sense of shame in their language, no table manners, here is 
Cybele 's foul *** and the freedom to speak in an effeminate voice. A crazed white-haired 
old man is the priest of the rites, a rare and memorable specimen of that enormous 
throat, an expert worth the fee. What are they waiting for? It's already time for them 
to use their knives to hack away their superfluous flesh in the Phrygian manner" (tr. 
S.M. Braund). As for the crux, fturpisf Cybeles is the reading of P,V and the contami- 
nated group O, except for two of the latter, which read turpes (cf. J. Willis ad loc. in the 
Teubner ed.). The old Loeb of G.G. Ramsey translated hopefully: "You will hear all 
the foul talk and squeaking tones of Cybele", which is obviously impossible. 

81 There are many parallel passages, e.g. Juvenal's description of a wild dinner-party, 
where the guests hurl drinking-cups at one another and the attendants because they 
get so drunk, as Corybantic, i.e. like those of the galli (5.24— 29). Jerome denounces 
the greed of Isiac and Metroac priests (Epist. 107.10). 



1 74 CHAPTER THREE 

bonds within the group and reinforcing their sense of sharing the same 
values. But seen from the outside, or critically, such exclusive feasts, 
annually repeated, might appear simply to be a means of openly 
reinforcing the gulf between those of a standing to take part in them 
and those who were not. Of course the elite was granted its exclusive 
privileges; it was when the custom moved downwards socially, among 
the priests and ordinary adherents of an exotic cult, that resentment 
was caused. 85 The ideological hostility such occasions aroused is thus 
hardly surprising 86 

Gluttony however is only one of Juvenal's accusations. There was also 
castration, always sure to arouse horror and disgust in the implied reader, 
as Juvenal suggests by his choice of the phrase supervacuam . . . carnem: 
(1 15). Castration of course implies feminisation [fracta voce); and femini- 
sation is equated with the pathic. 8 ' All are metaphors for the corruption 
inherent in the new. The intention is to tar Cybele herself with this 
brush, to make her, because she condones self-castration, responsible for 
all manner of aberrant sexual practices. The theme was greedily taken 
up by Christian invective, where the adherents of the Phrygian cults 
are a by-word for lasciviousness. One simply needs to cite a line from 
the Carmen contra paganos to see the sort of claim that was repeatedly 
made: Quern lasciva cohors, monstrum, comitaret ovantem (1. 66). 88 However, 
the indifference of Christian invective to truth or even plausibility is too 
familiar to require comment. The self-conscious hyperbole demanded 
of a Roman satirist has here been etiolated into a wearisome repetition 
of the claim to occupy the moral high ground. 

In fact, however, within the context of the Metroac concept of the 
divine world, Cybele defends chastity a outrance. What happens to Attis 
happens because he unilaterally breaks the requirement of sexual fidelity, 



85 Banquets on a considerable scale are known to have been held in the triclinium 
of the Serapeum at Ostia. The relative luxury of the priestly life must have been an 
important motive for the decision to join such groups, the human urge to fulfil spiritual 
needs being matched by the need to fulfil more immediately pressing ones. 

86 We might also mention here the accusation that the participants in the Phrygian 
cults make a fearful noise and fall into ecstasy (e.g. Rhianos ap. AnthPal 6.173; Diod. 
Sic. 3.57.8; Ovid, Fasti 4.243f etc., with Pachis 1996, 210-18). For some images of the 
instruments used in Metroac processions, see e.g. the r. and 1. lateral faces of CIL VI 
506/30782 = ILS 4414 = CCCA 3, no. 358 with pis. CCX and CCXII (c. 295 AD); 
cf. CIL XIV 385 = ILS 4162 = CCCA 3, 395; CCCA 3, no. 378 (both Ostia, IP). 

87 See Chap. 4.3. b below. 

88 See also Minucius Felix, Octav. 22.3-5; Arnobius, ^4<fo. nat. 1.35; 41-44; 2.5; 5.13 and 
16 etc.; Augustine, De civ. Dei 2.4.5; 2.6.7; 7.24-26; Prudentius, Pensteph. 10.10592 



SYSTEMS OF VALUE 175 

or restriction, laid upon him by the goddess. It is his breaking of the 
agreement that leads to the hasty self-castration (Ovid, Fasti 4.221-46). 
And the goddess' followers do likewise in order to avoid the danger of 
a similar temptation. 

We may now move on from the level of myth to the Claudian pseudo- 
history of the goddess' arrival in Rome. Here the theme of the goddess' 
concern with, and defence of, chastity re-appears in a good structuralist 
transformation. In this account, Claudia Quinta is the woman who 
invokes Cybele's aid when the ship carrying the sacred stone up to 
Rome gets stuck on a sand-bank in the Tiber. 89 Her chastity had been 
impugned just at the moment when all Rome was agog to greet the 
goddess from Pessinus. Whereupon the Great Mother duly intervenes 
in aid of the most chaste of Roman matrons. 90 Chastity was thus the 
first lesson Cybele taught the Romans. 91 Nothing could be further from 
the later accusations of lasciviousness. 

On the same theme, we may note the admiration evinced for those 
who were capable of remaining chaste even though they had not physi- 
cally gelded themselves. This is the implicit theme of Diodorus' account. 
In the context of the goddess' inventions in the realm of music, he 
praises Marsyas, who imitated her in inventing the double 'flute'; as an 
example of his intelligent self-discipline (sophrosyne), the historian men- 
tions that the proof of his moderation was his renunciation of sex: xr\q 
Se OKxppoo-uvrn; ot|lj£Tov eivoci <paoi to [ie%pi xr\q xeXevxr\q oOTeipaxov 
yeveo9ai xcov atppoStotcov (3.58.3). 

Moreover there are some examples of conduct by followers of Cybele 
irreproachable even in the eyes of the bienpensants whose views I quoted 



89 Ovid, Fasti 4.300-28; cf. Livy 29.10.4-1 1.8; 14.5-14; 36.36.3f.;Julian, Or. 5,160b. 
The bibliography on this 'event' is enormous; it will be enough to cite the contrasting 
views of Schmidt 1909, 1-18; Bomer 1964; Gruen 1990, 6 n. 3 ("Some troubling 
discrepancies exist between these texts, but do not require discussion here"); Borgeaud 
2004, 59-62, 69-71; Berneder 2004. See further, Chap. 4.3.a below. 

90 On the ethical implications of the Claudia Quinta episode, cf. Cosi 1986, 26: 
"[Gli fonti] non soltanto dimostrano l'esigenza nel culto di una generica purita rit- 
uale, ma testimoniano un insisto e significativo interesse da parte della dea per la 
castita". C.E. Schultz, Women's Religious Activity in the Roman Republic (Chapel Hill 2006) 
145 emphasises those versions that suggest that it was their unblemished reputations 
that led to the selection of both Claudia Quinta and P. Scipio Nasica to be on the 
welcoming committee. 

91 The second lesson must have been the importance of keeping agreements, shown 
by the triumph over Hannibal that occurred shortly after the goddess had been accepted 
into the pomerium (cf. Alvar 1994). This salvation in extremis could be interpreted as 
eloquent testimony to her saving power. 



176 CHAPTER THREE 

at the beginning of this section. If, instead of allowing themselves to be 
seduced by the tradition of moral invective, they had paid a little more 
attention to the epigraphic evidence, they would have found, among 
the mass of individual votives, a number of examples, mainly funerary, 
of conventional piety focused upon the family 92 From the late second 
century, the performance of the tauro-Zcriobolium, though undertaken as 
a public sacrifice for the good of the whole Empire, was also considered 
to be a means of individual catharsis, while in the late third and fourth 
centuries it might be seen as a statement of one's social and moral 
superiority 93 "In the minds of the devotees themselves, the public and 
private aspects of their religion were not clearly delimited" (Matthews 
1973, 178). One such late example is the altar recording the taurobolium 
and criobolium repeated at Rome by Ga[ma?]lios after an interval of 32 
years, no doubt sometime in the fourth century, which he describes as 
not merely a cruu|3oA,ov ex>TVxi"r\c,, a token of good fortune (1. 4), but 
as an offering consisting of his acts, his mind, his exceptional life, all 
the good things of his spirit: epyoc, voov, 7tpfji;iv, (3iov ei;o%ov, zaBXa 
nponavxa Ta[^.a]Xiov 7tpoc7uScDV (1. If). 94 

The most eloquent example of all however is again funerary. This 
is the joint monument, now in the Capitoline Museum, of the emi- 
nent senator Vettius Agorius Praetextatus and his wife Aconia Fabia 
Paulina dated c. AD 384. 95 In the course of her long address to her 
deceased husband, which is unfortunately too long to cite in extenso, 
Fabia Paulina writes: Te propter omnis me beatam me piam celebrant quod ipse 
me bonam disseminas totum per orbem: ignota noscor omnibus (rev, 11. 30-32). 96 
Although these texts are so late, there is in my view no reason to 
dismiss their implications for Metroac worship during the Principate. 



92 E.g. CIL VI 2260 = CCCA 3, no. 291; 2260 = ILS 4168 = CCCA 3, no. 292; XIV 
371 = CCCA 3, no. 423 = ASR 12.1 (Mythologische Sarkophage) 227 no. 76 (fine 
Alcestis sarcophagus from Ostia); XIV 408 = ILS 6179 = CCCA 3, no. 442; CCCA 3, 
no. 445 (both Portus); CIL XII 1567 - ILS 4140 = CCCA 5, no. 363 (Dea Augusta, 
Narb.); CIL III 2676/9707 = CCCA 6, no. 152; III 14243 = 158; 13903 = 167 (all 
Salona); CIL III 1 100 = ILS 7141 = CCCA 6, no. 485 (Apulum); IOSPE 1, no. 192 = 
CCCA 6, no. 517 (Olbia). However, there seem to be very few clear cases in Africa, in 
the western provinces, and in the entire Rhine-Danube area. 

93 On the history of the taurobolium, to which I return in 4.3c below, see Rutter 
1969 (better than Duthoy 1969); in the fourth century, McLynn 1996. 

94 IGUR HOf. no. 127 = CCCA 3, no. 239 (found beneath the Palazzo dei Conver- 
tendi in 1929). 

95 CIL VI 1779 = ILS 1259 = CCCA 3, no. 246, cf. Liebeschuetz 1999. 

96 "Because of you all salute me happy and religious, because you have broadcast 
my goodness to the world. I was unknown, (now) I am known to all." 



SYSTEMS OF VALUE 177 

Ordinary initiates, and not merely the elite of the galli, expected from 
their adherence to the cult a marked and continuing improvement in 
their ethical life. 97 

3. Isiac Ethics 



e7to|avu|iou 
[8e Koci o'vq npoGKX>v$> Q]eovq ax>vxt]pr\aeiv 
[kou <pvX6fyiv to 7iapa8]e8o|ieva u.01 pa)axr\- 

[pia 

Isiac oath from Oxyrhynchus: PSocItal 1 162 
= Totti no. 8b 11. 4-7 98 

My discussion of Cybele and Attis has suggested that the way forward 
is to examine the myth(s) for the implicit ethical values it contains. The 
situation in relation to the myth of Isis and Osiris is a little different, 
however, since there adulteries by husbands, suppressed by Plutarch, 
play an important symbolic role." Indeed, the narrative as a whole, 
with its incest, murder and violence, is unedifying It therefore seems 
best to use a wider range of information, especially in view of the fact 
that the sources are anyway more abundant. 

Let me start with a passage of Plutarch, which claims that it is how 
one approaches the deity that matters, as part of a commitment to an 
end through a rule of life: 

The name of her sanctuary, the Iseum, clearly offers recognition and 
knowledge of what really exists; for it is so called to indicate that we shall 



97 Cf. the inscription on the tabula ansata at the threshold of the Basilica Hilariana, 
probably a schola of the dendrophori of the Caelian (beyond the well-known image of 
the evil eye surrounded by aggressive animals and pierced by a spear): intrantibus hie 
deos propitios et basilic(ae) Hilarianae (CIL VI 30973a = ILS 3992, with the results of the 
later excavations of the site: A. Carignani, ArchLaz 10 (1990) 72ff; C. Pavolini, La 
topografia antica della sommita del Celio, MDAI(R) 100 (1993) 443-505). I take it that 
this divine care was felt to be earned. 

98 "I swear both to watch over the gods I worship and to keep secret the mysteries 
divulged to me." We possess two closely related versions of the Isiac oath, PSI 1 162 
(IIP) and 1290 (I p ), repr. as Totti nos. 8b and 8a respectively; both reveal the initiates' 
desire to fulfil the gods' commands, cf. R. Merkelbach, Der Eid der Isismysten, ^PE 
1 (1967) 55-73. 

99 Cf. Griffiths 1970, 316f. In the real Egypt, as opposed to the myths, adultery 
was severely punished: C. Eyre, Crime and Adultery in Ancient Egypt, JEA 70 (1984) 
92-105. 



178 CHAPTER THREE 

know what really exists if we approach the sanctuaries of the goddess 
with reason and reverence (\xexa A,6yot) kou 6o~i(0<;). 

De hide 2, 352a, tr. J.G. Griffiths (adapted) 

Both here and elsewhere (e.g. De hide 3, 352c), Plutarch is unfortunately 
not very explicit, and leaves the nature of what he would have consid- 
ered unethical behaviour, which no doubt includes magic, rather unclear. 
Nevertheless a good deal of relevant information can be extracted both 
from his essay on Isis and Osiris and from Apuleius' Metamorphoses. 

The first point to emphasise is one I have already made, namely 
that sexual abstention is one of the commonest moral injunctions in 
the cult. It may well be that this was a temporary abstention intended 
only to apply to the preparatory period before an important ritual. To 
cite Plutarch again: 

The process of consecration in the meantime, by means of a continu- 
ous and temperate regimen and abstinence from many foods and the 
pleasures of love (acbcppovi |a,ev evSeXexSk; 5iatTr| kou (3p(0|idt(ov noXkav 
kou d(ppo8tai(ov ct7toxai<;), keeps in check the unrestrained and pleasure- 
seeking element, and accustoms one to undertake austere and difficult 
services in sacred rites, of which the end is the knowledge of the First 
and the Lord . . . 

De hide 2, 351f-352a, tr. J.G. Griffiths 

Again, when Lucius is about to be initiated, he specifically alludes to 
the burdensomeness of the requirement of sexual abstention, castimo- 
niorum abstinentiam satis arduam {Met. 11.19.3). But such abstention was 
apparently also a condition of participation in one of the grand fes- 
tivals, something we hear about repeatedly in Augustan poetry. Thus 
Propertius complains: 

Tristia iam redeunt iterum sollemnia nobis: 

Cynthia iam nodes est operata decern. 
Atque utinam perant, Mlo quae sacra tepente 
misit matronis Inachis Ausoniis. ,wo 

2.33a. If. 
And again: 

Tu quoniam's, mea lux, magno dismissa periclo, 
munera Dianae debita redde choros, 



100 "Once again to my sorrow the dismal rites have returned: now for ten nights is 
Cynthia engaged in worship. Down with the rites which the daughter of Inachus has 
sent from the warm Nile to the matrons of Italy!" (tr. G.P. Goold). 



SYSTEMS OF VALUE 179 

redde etiam excubias divae nunc, ante iuvencae; 
votivas nodes et mihi solve decern! 101 

2.28.59-62 

The primary aim of such abstention, which was also practised in civic 
cult and other non-oriental mystery-cults, was to mark the special sta- 
tus of the deity and the gulf between human and divine. 102 It seems 
however to have been taken especially seriously in the cult of Isis. In 
Apuleius' Metamorphoses, the priest Mithras tells Lucius: 

Eat nunc et summo furore saeviat et crudelitati suae materiem quaerat aliam; nam in 
eos quorum sibi vitas in servitium deae nostrae maiestas vindicavit non habet locum 
casus infestus. 103 

Apuleius, Met. 11.15 

However, there can be no doubt that sexual abstinence was understood 
both as a mark of piety and as a technique to attain moral purity over 
a life-time. A couple of lines later, Mithras stresses that the goddess has 
guided Lucius to the point of being able to worship her. By becoming 
her slave, obeying her precepts, he will enjoy true freedom: teque iam nunc 
obsequio religionis nostrae dedica et ministerii iugum subi voluntarium, "dedicate 
yourself today to obedience to our cult and take on the voluntary yoke 
of her service". This same idea occurs in another passage that I cited 
in the previous chapter (p. 29), though it bears repeating here: 

But if by assiduous obedience, worshipful service, and determined celi- 
bacy you win the favour of my godhead, you will know that I — and I 
alone — can even prolong your life beyond the limits determined by your 
fate (ultra statutafato tuo spatia vitam . . .prorogare). 

1 1.6.7, tr. Hanson 

The priest Maiistas in the foundation-narrative of Sarapieion A at Delos 
proclaims that Isis and Serapis, as saving divinities, bestow benefits 



101 "My darling, since you have been released from mortal danger, perform the due 
service of a dance to Diana; perform also your vigil to her who, heifer once, is goddess 
now; and pay the ten nights pledged to me!" (tr. Goold). 

102 J. Bergman, Decern Mis diebus. Zum Sinn der Enthaltsamkeit bei der Mysterien- 
weihen im Isisbuch des Apuleius, in Studia G. Widengren, 1 (Leyden 1972) 332~46, has 
argued that these 10-day periods are to be linked with regular initiation and functioned 
as a sign of fidelity to the divinity, but, quite apart from Bergman's forced interpreta- 
tion of the ten-day period in relation to the Egyptian view that pregnancy lasted ten 
months, it seems odd to me that such a sign should precede the ritual death. 

103 "Let [Fortune] rage in all her fury and hunt some other object for her cruelty, 
for hostile chance has no opportunity against those whose lives the majesty of our 
goddess has emancipated into her own servitude" (tr. J.A. Hanson). 



1 80 CHAPTER THREE 

upon good people who have only pure thoughts in their hearts. 104 This 
is confirmed by one of the oracles given by Serapis to Timainetos, 
recorded in cod. Vindob. 130: 

6cyva<; X £ iP a ? £X«»v kcx! vow kou yA,(ottav &Xr|0f| 
eio<i>0i, |ir| A,oetpot<;, aXka voco KaSapoi;. 105 

Here the iteration of vovq is intended to make the contrast between 
the 'old' ex opere operato efficacy of standard purification and the new 
demand upon conscience quite unmistakable. The traditional indexi- 
calisation of the whole person, hands and mind, has been decisively 
slewed towards the latter. 

In this context there is a nice story told by Aelian about a cavalry- 
officer named Lenaeus who came to the temple of Serapis to have 
his valuable war-horse healed: it had been struck in the right eye and 
gone blind in it, and the large shield used by the cavalry blinkered the 
left eye, so he was useless (JVA 11.31). Lenaeus' main argument was 
that a horse cannot commit sacrilege (OeocruAacx) or murder; nor can 
it use impious language (|3A,occKpr|ula). Moreover, he himself had never 
wronged anyone (kou cxtjtoc; cbq o\)8e7ico7ioxe otjSevoc otjSev dSiicriacxc;). 
Evidently impressed by these moral arguments, the god attended to the 
animal, prescribing vapour baths at mid-day in the temenos instead of a 
fomentation. And it duly recovered its sight in that eye. 

The complement of such thinking was the interpretation of illness 
and misfortune as the result of divine anger. This is the specific appli- 
cation to the Egyptian deities of a very widespread belief in antiquity. 
Ovid briefly refers to such individuals seen in the streets of Rome: 

vidi ego linigerae numen violasse fatentem 

Isidis isiacos ante sederefocos. 
alter, ob huic similem privatus lumine culpam, 

clamabat media se meruisse via. 106 

Epist. ex Pont. 1.1.51-54 



101 IG XI.4 1299 = RICIS 202/0101 1.33f.: eo9XoTow 5e oacbxope^ aiev CTecGe I 
avSpdaw oi Kara 7idvra vocot oaia ippoveovoiv. 

105 "Enter with you hands and mind undefiled and your tongue truthful, pure not 
just by dint of lustrations but in your mind." Most conveniently to be found as Totti 
no. 61; the version in cod.Laur. 37 is slightly different. The date is uncertain, but prob- 
ably IP. 

106 "I have seen one who confessed to having outraged the godhead of linen-wear- 
ing Isis. He was sitting in front of her temple. And another who had been blinded for 
the same reason, crying out in the street that he deserved it." This and several other 
of these texts are cited by E Solmsen, his among the Greeks and Romans (Cambridge MA 
1979) 72. His comments however are surprisingly superficial, though I agree with his 



SYSTEMS OF VALUE 181 

What seems specific to the cult of Isis here, however, is the open con- 
fession of the cause of their misfortunes, presumably the result of a 
consultation of the priests. Such oracular pronouncements were thus 
the key institution linking misfortune and illness with the notion of 
specific personal wrong-doing, and thus led directly to rituals of pub- 
lic and private contrition, which were intended to purge adherents of 
their faults. Aristides states that Serapis grants a fitting destiny to all 
according to their deserts (Or. 45.25). It seems probable that the cult of 
Isis in the West was here picking up a theme of earlier Egyptian moral 
thinking, where Osiris was not the protector of all the dead but only 
of those who have attained a mode of life in accordance with Ma'at 
and managed to avoid the expansion of chaos in the world. 10 ' 

Moral desert is one of the central themes of Metamorphoses Book 1 1 . 
An important constituent of Lucius' sin was curiositas. 108 An additional 
reason for his being transformed was his enslavement to pleasure, ser- 
viles voluptates (11.15.1). 109 Lucius has to repent of all this if he wants 



general argument, that the cult of Isis become an integral part of the Graeco-Roman 
cultural superstructure. 

107 M. Zecchi, A Study of the Egyptian God Osiris Hemag (Imola 1996), on the story of 
Setne-Khamwas, dating from the end of the 2nd millennium (transl. by M. Lichtheim, 
Ancient Egyptian Literature, 3 [Berkeley and Los Angeles 1979] 125—51), which gives the 
impression that the protection of the good is a moralising inference from the belief 
in the judgement of the heart. I owe this information to Dr. M.A. Molinero. See also 
S. Quirke, Ancient Egyptian Religion (London 1992) 52ff 

108 A. Labhardt, Curiositas: Notes sur l'histoire d'un mot et d'une notion, Museum 
Helveticum 17 (1960) 206-224; R. Joly, Curiositas, L'Antiquite Classique 30 (1961) 33-44. 
The links between curiosity, ill-fortune and magic have been explored by S. Lancel, 
'Curiositas' et preoccupations spirituelles chez Apulee, RHR 160 (1961) 25-46, at 32ff ; 
J. Penwill, Slavish Pleasures and Profitless Curiosity: Fall and Redemption in Apuleius' 
Metamorphoses, Ramus 4 (1975) 49—82; Ch.-M. Ternes, De la metamorphose a l'initiation. 
L' ifineraire de Lucius dans les Metamorphoses d'Apulee", in Ries 1986, 363-376; PG. 
Walsh, The Rights and Wrongs of Curiosity (Plutarch to Augustine), Greece and Rome 
35 (1988) 73-85, urging a close relation between Augustine's Confessions and Apuleius' 
novel, though the new Augustine sermons provide surprisingly little support here, cf. 
F. Dolbeau, Nouveaux sermons de saint Augustin pour la conversion des parens et donatistes (IV), 
Recherches Augustiniennes 26 (1992) 69-141, cf. the transl. by E. Hill, Augustine of Hippo: 
Newly Discovered Sermons (New York 1997); also J. De Filippo, Curiositas and the Platonism 
of Apuleius' Golden Ass, AJPh 1 1 1 (1990) 47 1-492; P. Citati, La luce della notte, Materiali 
e discussioni per I'analisi dei testi classici 25 (1990) 165-177 at 171. LA. MacKay however 
adds a number of other sins, wealth, disobedience and the satisfaction of one's own 
desires as contributing to Lucius' punishment: The Sin of the Golden Ass, Arion 4 
(1965) 474—80. Hijmans 1995, 362—79 provides a systematic analysis of the theme of 
curiositas in the novel. 

109 Horace, Epist. 1.2.55f: sperne voluptates: nocet empta dolore voluptas, makes clear that 
such pleasure was considered addictive or enslaving. G.N. Sandy, Seniles voluptates in 
Apuleius' Metamorphoses, Phoenix 28 (1974) 234—244 links such pleasures to magic; cf. 
however Griffiths 1978, 155-58 on the higher voluptas; and Fick 1992, 273. 



182 CHAPTER THREE 

to obtain the goddess' grace again after his earlier life of dissolution. 
More or less subtle divine messages arrive, and at the end of his period 
of instruction, the initiand has learned his lesson. But he has to confess 
his faults and obtain the goddess' forgiveness before he is allowed to 
proceed to actual initiation. 110 

In connection with the negative connotations of sex in the novel, it 
may just be worth pointing out that there have been several attempts 
to read it in moralising terms, and that the sexual violence and disor- 
der of the first ten books disappear in the eleventh. 111 It is not simply 
that sexual relations are there condemned, but that the initiate actually 
becomes the goddess' consort (11.24). This symbolic status necessarily 
affects male-female sexual relations in the every-day world, since the 
underlying message is the protection of the organic family not founded 
upon the use of physical violence. "[Isis] represents sexuality within the 
orbit of the civilised community and the family: she represents marriage, 
and cannot condone illicit lust" (Griffiths 1978, 158). Chastity here is 
an ideal that can never be fully attained, yet it is precisely its Utopian 
status (in J.Z. Smith's terms) that allows it to act as a model for actual 
relationships in the world. 

The initiates of Isis were perfectly aware of the difficulties arising 
from attempting to live by the moral rules held up for them by the cult. 
In the heady excitement of initiation, they were convinced they could, 
by the goddess' aid, overcome their weaknesses. Some succeeded, as 
votives or funerary inscriptions occasionally indicate. One woman is 
called coniunx castissima, for example, another pientissima, a third coniunx 



110 There is a much later story about a senator who was briefly converted to Chris- 
tianity praying: "O goddess, I have sinned: pardon me. I have slipped back": Cyprian 
ap. CSEL 3, 302ff cited by Versnel 1990, 91 and n. 179. Despite the date, I find it 
interesting because it suggests something about the persistence of the notion of guilt in 
the cult of Isis. On confession in general, there is of course the vast compendium of 
R. Pettazzoni, La confessions dei peccati. 3 vols. (Bologna 1935), cf. the briefer summary, 
Confession of Sins and the Classics, HThR 30 (1937) 7if The sequence confession- 
judgement-forgiveness has been examined in detail by Griffiths 1982, which is more or 
less the same as idem 1991, 3132, cf. also Versnel 1990, 66 and 203ff 

in "'yjjg courS e of the narrative thus expresses a rejection of a phallic cult in favor 
of that of the maternal goddess": C.C. Schlam, Sex and Sanctity: The Relationship of 
Male and Female in the Metamorphoses, in Hijmans and Van der Paardt 1978, 95—105 
at 105; Winkler too stresses Lucius' moral development (e.g. 1985, 146). E. Finkelpearl, 
however, in line with today's dominant scepticism, urges: "Our conclusions about Lucius' 
moral development depend upon whom we believe: an ex-ass or a sophistic novelist": 
The Judgement of Lucius: Apuleius Metam. 10.29-34, ClassAnt 10 (1991) 221-36. 



SYSTEMS OF VALUE 183 

optima et sanctissima, bene de se merita. 112 Another man speaks of his wife, 
Flavia, who predeceased him, in the following terms: Primitiva, gratis- 
sima coniunxs, Flavia, et ipsa cultrix deae Phariae casta sedulaque et forma decore 
repleta, cum qua ter denos dulcissimos egerim annos.' 13 Although of course 
such epithets also belong to the conventional language of praise of a 
woman, we may perhaps take them, given what I have said about the 
wider Isiac context, as evidence that the cult was valued, among both 
men and women, precisely because it provided a structure both inside 
and outside the home for the acquisition and public demonstration of 
such virtues. 

Outside the home: that was precisely the problem for conventional 
morality, particularly in the caricatural form we find it in satire. One 
of the anxieties encouraged by the strict separation of male and female 
spheres is the fear of what women will get up to if left unsupervised at 
home (but even there, male slaves represented a danger) or are allowed 
out of the house by themselves, especially when there were strange 
men about. Such fears were very clearly at work in the imaginations 
of senators at the time of the Bacchanal affair in 186 BC. Insofar as 
the cult of Isis attracted women during the late Republic and well into 
the early Empire, so that they had occasion to 'meet' shaven-headed 
Egyptian priests in the temple, it is hardly surprising to find that such 
anxieties were again aroused. The cult was immediately suspected of 
encouraging vice. Ovid includes the temple of Isis in a list of places 
much used for assignations, while suggesting an elective affinity between 
goddess and illicit sex, via the familiar identification with Io: necfuge 
linigerae Memphitica templa iuvencae: multas illafacit, quodfuit ipsa Iovi (Ars 
amat. 1.78). 114 

Moreover, a famous case under Tiberius seemed to give colour to such 
fears and accusations; it certainly produced a reaction very similar to 
the case of the Bacchanals: the statue of Isis was thrown into the Tiber, 
the temple razed to the ground, the ceremonies forbidden, and several 
priests were crucified. The occasion was the debauchery of a chaste 



112 ILS 9442 = RICIS 512/0101 (which I take with Festugiere to be Isiac; Ravenna); 
CIL VI 34776 = RICIS 501/0161; 2249 = 501/ 0162 (both Rome). 

113 "Flavia Primitiva, my dearest wife, who was herself a chaste and devoted wor- 
shipper of Isis, whose beauty was (ever) proper, and with whom I lived thirty extremely 
happy years": CIL VI 17985a = RICIS 501/0177 (sarcophagus from Rome, perhaps 
mid-HP). 

114 Loosely: "Do not avoid the Egyptian shrine of Io clothed in linen; she makes 
sure that plenty of girls end up doing what she did with Jupiter." 



1 84 CHAPTER THREE 

married woman, Paulina, by Decius Mundus, a man of equestrian rank 
who had fallen in love with her and been several times sent off with a 
flea in his ear. He had even offered her a large sum of money to oblige 
his infatuation. According to the highly, perhaps suspiciously, circum- 
stantial account provided byjosephus (Ant. Jud. 18.65-80), Paulina was 
very much devoted to the worship of Isis: 9epa7ieioc xr\c, "IcnSoq atpoSpoc 
■{)7rr|Yuevr|v (§70). Mundus' freedwoman Ida bribed several of the priests 
of Isis to find a means of satisfying Mundus' passion. One of them 
told Paulina that she had been chosen to share the bed of Anubis. In 
this belief she duly went to the temple, attended by the priests, and 
there had intercourse with Mundus alias Anubis. Her mouth enlarged 
by the belief that she had encountered the god, she made the most 
of the story in her social circle, until Mundus, unable to contain his 
triumph, told her the truth. The husband informed the emperor, who 
reacted in the manner I have already described. 

It hardly matters whether, or how much of, this story is true. 115 All 
that matters is that many people believed it, and that it served as the 
official reason for persecuting the cult: although the Iseum claimed to 
be the temple of chastity, this was not in fact the case. It was this les- 
son that the moral majority learned. From that point of view, Juvenal's 
epithet lena for Isis was not an outrageous slur but a simple fact. 116 

These were not the only grounds for Roman ambivalence about the 
cult. The elegiac poets complain at being compelled to sexual abstinence 
while their lovers are away at the Iseum for days at a time. The cult thus 



115 It has been plausibly suggested that this was a political persecution, though there 
is no direct evidence for it, cf. Dunand 1980, 7 Iff Takacs 1995 is surprisingly super- 
ficial here, since, though she agrees with Dunand, she does not seem to realise that 
conclusions drawn from a single type of evidence may distort reality. She thus argues 
that the numismatic evidence for the late Republic shows that the cult was not revo- 
lutionary as A. Alfoldi, Isiskult und Umsturzbewegung im letzten Jhdt. der romischen 
Republik, Schweizerischer Miinzblatter 5 (1954) 25—3 1 once claimed (whose arguments were 
dismissed already by Crawford, RRC p. 584 no. 2), without making the obvious point 
that evidence of this type could never prove anything relevant to such issues. If she 
had taken the literary evidence, essentially the elegists, into consideration, she would 
perhaps in this case have qualified her point — for which there is otherwise plenty to be 
said — that the Senate's repression of the cult was less ethical than political. In my view 
Malaise's discussion of the role of political considerations in the cult's expansion is the 
most successful (1972a, 357ff). W.C. van Unnink, Flavius Josephus and the Mysteries, 
in Vermaseren 1979, 244—79 at 278, argues that we should concentrate less on the 
incident itself than on the general image of the mysteries purveyed byjosephus, a Jew 
convinced of the superiority of his own religion, in the Flavian period. 

116 The wife is imagined as dolling herself up for assignations with a lover in hortis 
aut aput Isiacae potius sacraria lenae (Sat. 6.488f). 



SYSTEMS OF VALUE 185 

interfered with their own sexual pleasure, provided a sphere in which 
women could be temporarily independent of someone else's sexual 
will. At the same time, however, we can use this evidence to reinforce 
the general claim that the cult of Isis genuinely did place a high value 
on chastity and sexual abstinence. It never occurs to these poets, for 
example, to imagine that their muses were not spending their ten days 
of chastity honestly. If they had had any suspicions, we can be sure 
they would have expressed themselves differently 117 Tibullus imagines 
Delia, dressed in pure linen, sitting awake at night in the temple: 

nunc dea nunc succurre mihi . . . 
ut mea votivas persolvens Delia nodes 
ante sacras lini tecta fores sedeat. . . lls 

1.3.27-30 

Propertius too: 

at tu, quae nostro, nimium pia, causa dolori's, 
noctibus his vacui, terfaciamus iter™ 

2.33a.21f. 

All this, to my mind, gives colour to the topos that is central to several 
of the novels, that it is Isis who protects the heroine's chastity. We may 
think of Leucippe in Achilles Tatius' Leucippe and Cleitophon, Anthea 
in Xenophon's Ephesian Tale, or, best of all, Chariclea in Heliodorus' 
Aethiopica. m At the same time, we have to accept, as I have already 
argued, that there may often have been a decalage between aspiration 



117 J. Alvar, Marginalidad e integration en los cultos mistericos, in F. Gasco and 
J. Alvar (eds.), Heterodoxos, reformadores y marginados en la Antigiiedad Cldsica (Seville, 1991) 

71-90 at 72f. 

118 "Now aid me, goddess. . .that my Delia may watch through the night, in accor- 
dance with her vow, sitting all dressed in linen before your holy doors . . .". 

119 "But you, who through an excess of piety have caused my sufferings, when we 
are released from these nights, let us thrice make love's journey" (tr. J. Goold). Goold's 
text makes better sense than Barber's (OCT 2 ), reading nimium placata dolore es. 

120 The more recent literature includes I. Stark, Religiose Elemente im antiken 
Roman, H. Kuch (ed.), Der antike Roman: Untersuchungen zur literarischen Kommunikation 
und Galtungsgeschichte (Berlin 1989) 135-149; S. Wiersma, The Ancient Greek Novel 
and its Heroines: a Female Paradox, Mnemosyne 41 (1990) 109—123; Konstan 1994; 
R. Merkelbach, Novel and Aretalogy, in Tatum 1994, 283-95; Beck 1996. On this 
particular point note too: J. Hani, Le personnage de Charicleia dans les Ethiopiques: 
Incarnation de l'ideal moral et religieux d'une epoque, Bulletin de I' Association Guillaume 
Bude 1978, 268—273; E Napolitano, Leucippe nel Romanzo di Achille Tazio, Annali 
delta Facoltd di Lettere e Filosqfia delta Univ. di Mapoli 26 (1983—84) 85—101. 



1 86 CHAPTER THREE 

and reality, between the resolutions and commitments made at religious 
high-points, and the reality one found oneself living each day 

In the case of the Egyptian cults, we happen to possess a quite dif- 
ferent source of evidence for the value-system, namely the aretalogies, 
the 'self-predications', the 'hymns'. None of the other cults can offer 
anything comparable. A number of related types of document existed 
in antiquity. Perhaps the most interesting for my purposes are the five 
surviving versions of a text evidently composed in Greek at Memphis in 
the Fayyum around 1 00 BC or somewhat earlier, and generally, though 
perhaps somewhat incorrectly, known as aretalogies of Isis. 121 There is 
little point here in getting bogged down in this issue. 122 My own view 
is that the traditional term, although it does not accurately capture the 
entire content of these texts, has the advantage over other suggestions 
of accurately communicating their general purpose. 123 



121 I can dispense with references to the older literature because the documents 
are now available, with good texts, full bibliographies in RICIS (with a useful French 
translation) and/or Totti 1985. It should be noted that I group nos. 1-5 together as 
representatives of the Urtext at Memphis, but in fact only nos. 3—5 are identical, and 
nos. 1 and 2, i.e. the earliest versions, diverge in many important respects from the later 
standard text: 1) Maroneia (late IP ): RICIS 114/0202 = Totti no. 19; 2) Andros (P): 
RICIS 202/1801 = Totti no. 2; 3) Kyme (FTP): 302/0204 - Totti no. la (a composite 
text combining Kyme, Thessalonike, Ios and Diodorus); 4) Thessalonike (I— II 1 ' ): RICIS 
113/0545 (fragment); 5) Ios (IIP): RICIS 202/1 lOl(fragment). To these we may add 
some related texts: the prayers from Kios in Bithynia (I p ) (RICIS 308/0302 — Totti no. 
5) and Cyrene (AD 103) (RICIS 701/0103 = Totti no. 4); the long self-predication of 
Isis in POxy. 1380 = Totti no. 20; the self-predication at the end of the Kore Kosmou 
(ap. Stobaeus 1.49.44 = Festugiere and Nock, Corpus Hermeticum 4, frg. XXIII = Totti 
no. 3); the four Hymns by Isidoros, edited Vanderlip 1972 — Totti nos. 21-24 = Sanzi 
2003, 206-13, Isis nos. 45.1-4 (P); cf. the translation by M. Gustafson in Kiley 1997, 
155—58; and the hymns contained in a number of literary texts: e.g. Diodorus 1.27 
(Bergman 1968, 27-43 = Totti no. la,b); Tibullus 1.7.29-48 (= Totti no. 7, cf. Alfonsi 
1968; Koenen 1976); Apuleius, Met. 11.2; 5-6; 25.1-6. We could add some others, 
such as Aristides' so-called Hymn to Serapis (Or. 45) and a couple of minor texts of 
less relevance to the topic here. 

122 Longo 1969, 47-52 wanted to distinguish between sacred aretalogies and pro- 
fane ones. If aretai are necessarily 'actions', divine ones must be too, so that the term 
aretalogy is appropriate; however the actual content of these texts goes far beyond the 
semantic field of aretai, so we should avoid the term. Grandjean 1975 used the word 
aretalogy quite unself-consciously. Henrichs 1984 wished to retain the traditional term 
on the grounds that the aretalogies' resemblance to one another nullifies the spontaneity 
of prayer. Totti 1985 introduced a number of different terms: Selbstoffenbarung, Hymnus, 
Preis, Aretalogie, Enkomium without clearly distinguishing them; through trying to show 
awareness of the various sub-genres involved, Bricault avoids many of the traditional 
terms in his entries in RICIS, which has led to a number of problems in the indices, 
cf. R.L. Gordon in CR 57.1 (2007) 232-4. 

123 P. Roussel, Un nouveau hymne a Isis, REG 42 (1929) 137-68 and W. Peek, Der 
Isishymnos von Andros (Berlin 1930) both rejected the then traditional word 'hymn', cf. 



SYSTEMS OF VALUE 187 

The fact that there is a range of analogous documents all devoted 
to praising one or other of the Isiac group of deities produced a con- 
siderable debate about their origins in the older literature. The various 
possible positions have been refined and improved in the course of this 
debate, but no consensus has been reached. Richard Harder argued 
in 1944 that these texts were written within the context of Egyptian 
religiosity. 124 After the war, this was disputed by A.D. Nock and Le 
Pere Festugiere, both of them Hellenists, but also in the late 1950s by 
the Egyptologist D. Miiller. 123 The Hellenists seemed to be gaining the 
upper hand, but then the Swedish Egyptologist Jan Bergman intervened 
strongly in favour of an Egyptian background (Bergman 1968). Yves 
Grandjean, the editor of the earliest known version, from Maroneia, 
which explicitly identifies Isis with Demeter, was unable to rebut his 
arguments (Grandjean 1975). Nowadays, however, as so often with these 
older disputes, the entire debate seems pointless, since it has become 
clear, that, just as in the analogous cases of the Hermetic literature 
and the magical papyri, or even astrology, there can be no question 
of an either/ or answer: these are all products of a complex process 
of cultural translation and symbiosis that affected both sides in equal 
measure, so that it makes little sense to weigh this thought against that 
and wave it as 'proof of anything. 

As I have made amply clear, the oriental cults of the Roman period 
were in fact a complex bricolage of elements drawn from high cultures 
notionally 'other' than Graeco-Roman, elements in each case irrevers- 
ibly, even unrecognisably, conditioned by conquest, fragmentation, 
translation and reception, a reception that required, among other 
things, that they measure up to the implicit demands of Eleusinian 
cult. What is important therefore is not what they were, that is, had 
been, at some arbitrary point in the past, but how, why and where they 
were received, and the preconceptions in terms of which they could be 



Miiller 1961, 15; E. Norden, Agnostos Theos (Stuttgart 1912/1966) 149ff; Festugiere 
1949. The term 'prayer' is very vague and marginalises many important features of the 
true aretalogies. I. Cazzaniga, Intorno ad alcuni epiteti di Iside nella litania di P.Oxy. 
1380, Parola del Passato 20 (1965) 23311 once suggested the word 'litany'; O. Weinreich 
rather portentously offered 'Res Gestae Reginae Isidos', which of course never caught 
on: Antike Heilungswunder (Giessen 1909) 46 n. 7. 

124 R. Harder, Karpokrates von Chalkis un die memphitische Isispropaganda (Berlin 1944). 

125 A.D. Nock, Gnomon 21 (1949) 221-28 and Festugiere 1949 both objected that all 
the relevant terms were Greek; Miiller 1961 insisted on the degree of transformation 
that Egyptian Isiac themes had undergone in the transition to the Hellenistic world, 
a well-taken point. 



1 88 CHAPTER THREE 

regarded as cultural goods of a specific type of value. 126 With regard 
to the Egyptian cults, as we have seen, we happen to know from Plu- 
tarch that Manetho the Egyptian priest and Timotheus the Eleusinian 
co-operated in the creation of the cult of Sarapis (p. 59). He was to 
be Isis' new companion during the Hellenistic period, which provided 
these two master-alchemists with the opportunity to create an entire 
sub-set of new rituals. I would say it is no coincidence that we possess 
no aretalogies dating from earlier than the late second-century BC: 
there probably were none before that time, because they are a type of 
religious expression that was inspired and made possible by their joint 
work. We cannot of course tell whether they created a proto-aretalogy 
since there is no relevant evidence; however I can certainly imagine 
that they were not content with merely drawing up a sketch-plan but 
that they got down to drafting an entire system of beliefs and rituals 
appropriate to their syncretistic creation — which would mean that 
there is no sense at all in asking whether the aretalogies are 'basically' 
Greek or Egyptian. 

If that is so with regard to their form, the case is similar with regard 
to their religious value. They seem to me to be lists of virtuous actions 
worthy of reverence but also of emulation. They may be lists of divine 
acts, but their followers are invited to follow their example, so far as 
they can. In other words, we may properly call part of their content a 
set of ethical norms. It is precisely their status as hymns that confers 
upon them an undeniably didactic quality: they are not so much prayers 
as chanted mnemonics whose function is to transmit the value-system 
envisaged by the cult. They were probably not intended originally to be 
learned individually — there are too few copies for that — but to be recited 
collectively and thus learned by heart through constant repetition. The 
copies we possess were all votives intended to add distinction to the 
furnishings of the temple; and anyway only a few people could even 
read inscribed texts in Greek. No comparable inscriptions have been 
found in the Latin-speaking West. We therefore do not know whether 
there were translations of them into Latin, nor if there were analogous 
rituals there of collective chanting — at Rome we hear of a young man 
who held the post of aretalogus, but it is specifically said to be for texts 



Cf. J. Alvar, Isis preromana, Isis romana, in Rubio 1997, 95—107. 



SYSTEMS OF VALUE 1 89 

in Greek. 12 ' At any rate, the value-system surely did spread in the West 
too, and we have every reason to think that the contents of the aretalo- 
gies were intended to have universal or general application. 

Ilap' e|ioi to Sikouov ia/vei, declares Isis, "With me justice prevails" 
(Kyme §38). 128 The Maroneia aretalogy claims that Isis created justice 
so that mankind may live in equality just as we all alike, in the course 
of nature, must die (11. 24-26: a\>xt\ to Sikcxiov eoTT|oev iv' ekcxotoc; 
filicov, c&q 8K Tilt; (pvoecoq t6v BdvcxTov i'oov eo%ev, koc! ^fjv anb tcov 
i'ocov eiSfj). 129 And so that they may live with one another and com- 
municate, not simply men with women but everyone with everyone 
else, she instituted language, foreign languages for foreigners, Greek 
for the Greeks (11. 26-28: avxt] tcov dv9pamcov oi<; |iev pdppapov, 
oic, Se eAAriviScx SidA,£KTov earnaev, iv' f)i to yevoc; SiaAAdaaov |xr) 
|i6vov dvSpdatv npbc, yovaiiccxi; dA-A-oc Kal naai npbq 7TdvTocq). 130 She 
has bestowed laws, so that all the cities may flourish, since they are 
ruled not by violence with legal backing but by law without violence 
(11. 29-31: [a]x> v6|ioi)c; eScoKcx<;...TOi[ycx]po'6v od noXzic, e\)OTd9r|oav, 
ov ttiv (3(av vo|iikov dA,A,d [t]6v v6|j.ov dpiaoTov evpovoai). 131 These 
elements of the doxology all imply correlative behaviour on the part 
of worshippers. Moreover, telling the truth, and the fairness of the 
judicial system, are of great importance (Kyme §28f; 37). As for the 
institution of marriage, the Maroneia text just refers to the union of 
Isis and Serapis (1. 17), implying its model function, adding later that 
children should honour their parents as though they were very gods, not 
simply fathers (11. 31-33: cxo Ti|ida0ou yoveiq vnb [t]ekvcov znoi~r\aac„ 
ou |iovov &>q 7taTepcov, dA,A,' &>c, koc! Oecov [tpJpovTiaocaa). 132 The Kyme 
self-predication, however, shows much greater interest in the entire 



127 AE 1999: 349 = RICIS 501/0214, with the remarks of M.P. Del Moro, Via 
Latina 135: cronaca di un intervento di urgenza, RAC 75 (1999) 33-36. Presumably it 
was normally the duty of one of the priests to recite these texts. 

128 Cf. Kyme §16: eyd) to 8iKaiov ioxupov enoirioa. 

129 There is no correlate in the Kyme version. 

150 Cf. Kyme §31; in Maroneia 22~ 24, Isis invents writing, together with Toth- 
Hermes, icai xSv ypa|X|j.dTO)v a |iev iepa toTi; iivaxaiq, a §e Snuoaia xoti; nactv, thus 
showing where the true priorities lie (in Kyme §3c this distinction is made in terms of 
hieroglyphs versus Demotic). 

131 The Kyme version distributes aspects of what we might call cosmic and social 
justice over different sections (§§12-14, 16-21, 25f, 28-30, 36-38, 40, 52), and names 
a number of specific crimes and abuses (tyranny, massacres, false oaths, legal entrap- 
ment, fraud, ignoring the rights of suppliants); cf. Kore Kosmou 1. 6f. Totti. 

B2 'T^g same theme appears in Kyme §19. 



190 CHAPTER THREE 

theme. Isis declares for example: eyro yovouKOc koc1 6cv8poc covnyocyov 
(§17); eycb owypoctpocc; yocuticac; evpov §30; eyro axepyeaBou yuvouiccxc; vnb 
ocvSprov f)vdyKaoa §27; eyro xdic, 6caxopy<ro>c; yovevow Stcxiceiuevotc; 
xeturo[p]{ocv e7xe9r|Ka §20. 133 The Egyptian gods in fact maintained the 
fundamental values of the community, through the institution of the 
family based on asymmetrical heterosexual marriage, and regulated by 
a system of justice whose rules are set out in writing. Thanks to Isis, 
society has evolved beyond the early stages, where discord was typical. 
Harmony is thus a good, but even to make that claim is to concede 
that conflicts are going to persist: eyro xovq StKcuroq cVwouivoi)!; xeturo, 
"I honour those who defend themselves when they are in the right" 
(Kyme §37). 

This structure of opposites reveals the extent to which the Isiac 
order sustained and supported the established order of things: o\)9ei<; 
So^d^exai avzv xfjc; eufjt; yvrouriq, "no one rises high without my 
assent". This is not so much because it was in itself reactionary as 
because those who composed and edited the aretalogies inevitably 
also had the opportunity of instrumentalising them in the interests 
of a conservative view of the natural and proper order of society. Isis 
offered a means of sublimating social asymmetry without herself will- 
ing it. As a deity, she is herself beyond political, social or economic 
action, but she is responsible for the fact that things are as they are 
and not otherwise. Indeed she claims that good and bad conduct, and 
by implication the social structure that sustains actual moral codes, 
are matters not of social convention but of Nature: eyro xo kocA,6v kcxI 
odaxpo[v] StcxyewroaiceaGou vnb xr\q (pvoeroq e7xovr|acx (Kyme §32; cf. 
Maroneia 1. 33f). To that extent those who enjoy positions of power 
are exonerated from responsibility so long as they do not actually go 
against the law; their privileges, and conversely the sufferings of the 
wretched, proceed from divine dispensation. In other words, there is 
no point in looking to the cult of Isis for social change. 134 



133 "I have caused women and men to lie together (in marriage)"; "I have invented 
marriage-contracts"; "I have laid it down as a natural law that women should be 
cherished by their husbands"; "I have decreed punishments for parents who do not 
cherish their children". 

134 The aretalogical theme of Isis' dominion makes her elective affinity with the 
powers that be quite plain. She may break the chains of destiny but she herself is 
also a 'tyrant' (b/m euii f| ropavvoi; 7cd<yr|(; x^po-i, Kyme §3a) Henrichs 1984, 155-57 
and Versnel 1990, 39—95 (with an analysis of the term 'tyranny' in relation Isis) have 



SYSTEMS OF VALUE 191 

It has been claimed that, despite being at the service of the estab- 
lished order, the cult encouraged a separate sphere for women, while 
never questioning their inferiority to men. Whether women did in fact 
use this opportunity to reduce their dependence on men in however 
unclear. 135 The very literary aretalogy from Andros puts the point quite 
bluntly: Acokcx x«p[i^6|ieva] 7iei0dviov dvSpi ywomccx <V(poTepco[<; x' 
eSdncxaacx. .., "I have kindly given to husbands obedient wives, and 
subjected them both to the yoke". It would not be too cynical to see 
here a main reason for the appeal of the Egyptian cults to men: like 
Islam in West Africa, they offered to bolster the unstable authority of 
married men. It is even possible that the cult was of value as a means 
of encouraging 'positive dependence': from the early Principate the 
powers that be and the leaders of the Egyptian cults seem to have 
connived at turning the latter into an element of Roman culture that 
could effectively help integrate provincial populations into Romanitas. 
This would explain the high rate of participation by men, and avoid 
the need to postulate, as some models do, a male version of the cult 
and a female one (Kraemer 1992, 77). It is clear of course that both 
gender and status were important. That is, people's expectations of 
the cult varied in relation to such differences, but that is quite different 
from claiming that there were different cults. I would only say that the 
priests' version was substantially different from that of ordinary, less- 
or uneducated initiates. The difference here however lay not so much 
in the praxis itself as in their contrasting perceptions of it. Moreover, 
the fact that people's anxieties and needs differed, inter alia down the 
gender divide, facilitated the emergence of contradictory discourses 
within the cult of Isis. After all, the cult proposed an omnipotent god- 
dess who was nevertheless always accompanied by a male (Alvar 1999): 
this sets up one model, of a dependent spouse happy at home, but at 



recently emphasised the political implications of the cult, which helps us to understand, 
to an extent, its attraction for men. 

135 The role of women in the cult, and their motives for joining it, have been much 
discussed of late. Many scholars have argued that the values it proclaimed, though 
overtly integrative, did encourage a certain autonomy and self-wareness: Heyob 1975; 
Dunand 1980; J. Alvar, Las mujeres y los misterios en Hispania, Actas de las Quintas 
Jornadas de Investigation Interdisciplinaria: La mujer en el mundo antiguo (Madrid 1986) 245-257; 
idem, La mujer y los cultos mistericos: marginacion e integration, in M'\J. Rodriguez, 
E. Hidalgo and C.G. Wagner (eds.), Jornadas sobre roles sexuales: La mujer en la historiay 
la cultura, Madrid, 16-22 mayo 1990 (Madrid 1994) 73-84; Mora 1990. Kraemer 1992, 
71-79, who depends heavily on Heyob and Dunand, stresses the cult's ambiguity 
between traditionalism and innovation. 



192 CHAPTER THREE 

the same time another, that of a more emancipated and autonomous 
woman. Thus, despite the Andros text, two of the claims of the Kyme 
aretalogy seem to point in the same direction as the image of equal 
marriage proposed in the novels. 136 

There is one last point to make about the position of the Egyptian 
gods in the Roman world. To judge from the number of surviving 
bronze and terracotta statuettes, both Isis and Osiris seem quickly to 
have established themselves among the private deities of the household, 
sharing the lararium with the other gods specially favoured by each fam- 
ily 137 It has been suggested that Osiris was adopted into this sphere as a 
result of the journey of Tibullus' patron, M. Valerius Corvinus Messala, 
to the Delta (cf. Tibullus 1.7.23-54). In this passage, Osiris appears as 
the protector of the family store-room, a god of agriculture, alongside 
the Penates and Vesta, but also as the god of the deceased members 
of the family and thus, like the Lares, crucial to the preservation of the 
Roman family through time. All this already in 28 BC. 138 It has also 
been remarked that, at any rate in the western Empire, there seems to 
have been no attempt to subject the Egyptian gods to further, 'Celtic' 
or 'Germanic', interpretation. 139 In the north-western provinces, they 
retained their difference, their distinctive character. 

4. Moral Values in Mithraism 

Now when the watch is ended, now when the wine 
is drawn, Mithras, also a soldier, keep us pure till 
the dawn! 

R. Kipling, A Song to Mithras (from Puck of Book's 

Hill, 1906) 



136 Cf. n. 120; also Kraemer 1992, 77; Konstan 1994. 

13/ This no doubt had something to do with Isis' traditional role as KorjpOTpocpoi;, lac- 
tans, mother of Horus, cf. Muller 1969b; Tran tam Tinh 1973; Malaise 2005, 151f. 

138 Qf Turcan 1996a, 90; Takacs 1995, 79. The phenomenon is not confined to 
Roma and Ostia (cf. Bakker 1994, 41). Provincials probably took over the Egyptian 
gods into their lararia in the same way: small statuettes of these divinities, in bronze 
and terracotta, are after all widely disseminated over the Roman world (cf Mostra hide, 
264-71 nos. IV 290-304; 571-73 nos. VI 33-38). 

139 U.-M. Liertz, Isis und ihre nordische Schwester, Arctos 37 (2003) 101—14; this 
applies even in the case of the three examples of Isis Noreia (AE 1954: 98 = ILLPROJS 
1 181; CIL III 4809 = ILLPRON 151; EpDatHeid 014591). 



SYSTEMS OF VALUE 193 

The absence of a proper myth of Mithras means that we cannot anal- 
yse the cult's moral claims in the same manner as in these two cases. 
Moreover, the literary references are so few, and so tangential to this 
theme, that we cannot make much headway there either. These difficul- 
ties ought not however to discourage us, or induce us to imagine that 
the cult possessed no moral system. 140 One clear indication of this, and 
suggestive of a possible thematic continuity from Iranian culture into 
the Graeco-Roman world, is the frequent representation of Mithras 
shaking hands with Helios/Sol, which combines the idea of a contract 
or agreement with that of acceptance into a group of friends. 141 The 
gesture occurs repeatedly in Roman iconography to convey the moment 
of solemnisation of agreements, treaties, pacts, and also the expression 
of friendship, loyalty and alliance, but was also central to the Iranian 
conception of MiOra. 142 

Less certain evidence of such continuity in the ethical sphere is the 
theme of justice. 143 We do know of one inscription in Asia Minor where 
Mithras has the epithet SiKouoq. 144 This however is almost certainly due 
to his association with the Sun, the theme of the Sun's justice being 
both ancient and widespread in Anatolia. To be sure, the fifth-century 
BC MiQra-Yast leaves no doubt of the close link between Iranian MiOra 
and the maintenance of justice, which is of course inherent in the idea 
of the sanctity of agreements and contracts; and in Iranian tradition 
he was one of the three judges of the dead at the Cinvat Bridge. 145 
Gwyn Griffiths has argued that this concept of Mithras as judge was 
also present in the Roman cult, though the evidence for it is admittedly 



"" Cf. the arguments of J.H.W.G. Liebeschuetz, The Expansion of Mithraism 
among the Religious Cults of the Second Century, in Hinnells 1994, 195-216 at 197, 
in favour of the claim that Mithraism had a well-developed moral system and offered 
forgiveness of sins. 

111 Cf. Rostovtzeff et al. 1939, 122; LeGlay 1978; Merkelbach 1984, 61; Clauss 
2000, 15 If. 

112 On miOra- as the Avestan common noun for 'contract', see Gershevitch 1959, 
26-30; cf. P. Thieme, The Concept of Mitra in Aryan Belief, in Hinnells 1975, 21-39; 
Merkelbach 1984, 4f 

11! This was an important factor for Cumont, e.g. 1929, 144: "Mithra . . . est reste 
en Occident le dieu de la verite et la justice". 

114 V. 18 (Tyana in Cappodocia). The text, which consists of the three words: QeS> 
SiKaicp Mi0pa, is however as uninformative as might be. 

145 Cumont 1929, 147; Merkelbach 1984, 23f; Griffiths 1982, 209f.; 1991, 313ff 



1 94 CHAPTER THREE 

thin to the point of non-existence. 146 The initiatory scenes painted 
on the podia of the mithraeum at Capua near Naples can hardly be 
used to support the idea that punishment was one of the themes in 
Mithraic initiation. 14 ' The complete silence of the epigraphy suggests 
that Mithras' justice was not an especially important aspect in the 
western cult, if it existed at all. 

Although it is good method to be sceptical with regard to Christian 
claims about the oriental cults, in this area as in others, a passage 
by Justin Martyr, dating from the mid-second century, alludes to the 
emphasis placed by Mithraists on 8iiccxi07tpcxi;{cx and itself cites a pas- 
sage from Isaiah: 

Now, when those who hand down the mysteries of Mithras claim that he 
was born of a rock, and call the place where they initiate his believers 
a cave, am I not right in concluding that they have imitated that dictum 
of Daniel, "a stone was cut without hands out of a great mountain" [Dan. 2.34]? 
In similar fashion, have they not attempted to imitate all the sayings of 
Isaiah? For the demons urged the priests of Mithras to exhort their fol- 
lowers to perform righteous acts. [2] Here are the words of Isaiah which 
I must quote that you may know from them that this is so: "Hear, you who 
are far off what I have done; and they who are near will know My strength. The 
sinners in Zjon have departed; trembling will seize the impious. Who will announce 
to you the everlasting place? He who walks righteously, and speaks truth, who hates 
iniquity and injustice, and keeps his hands clean from bribes; who stops his ears from 
hearing the unjust judgment of blood, who shuts his eyes from seeing evil; he will 
dwell in the lofty cave of the strong rock. [3] Bread will be given to him, and his 
water will be sure. You will see the King in his glory, and your eyes will look far off 
Your soul will meditate on the fear of the Lord. Where is the scribe? Where are the 
counselors? Where is he who counts those who are nourished, the small and the great 
people? With whom they did not take counsel, nor knew the depth of the voices, so 
that they did not hear. A shameless people, and there is no understanding in him who 
hears" [Isaiah 3343-19]. 

Dial. Tryph. 70.1—3, tr. Falls/Marcovich 

It might be claimed that Justin is projecting onto the Mithraists moral 
demands that were in fact characteristic of Judaism or Christianity, so 
that he can claim that they are the result of diabolic imitation, just 
as he does in relation to the eucharist: orcep kocI ev xdiq xov MiBpoc 
rrocruripioK; TCapeScoiccxv y(veo9ai ruuriadnevoi oi 7tovr|poi Saiuovec, ... in 



we 'pjjg evidence is so poor that Griffiths, never very discriminating in his choice 
of evidence in these matters, is even reduced to citing HA Commod. 9.6: sacra Mithraica 
homicidio vero polluit. 

"' See the discussion below, Chap. 4.5.b. 



SYSTEMS OF VALUE 195 

the passage from the Apology. Against that, the theme of Siicocio7tpcxi;{oc 
occurs in no other Christian writer, and is extraneous to the well-known 
facts that Justin also cites about Mithras' birth and the word for the 
mithraeum. And in fact the most direct evidence we possess about the 
ethical requirements of Mithraism, a passage from Porphyry's De antro, 
tends to confirm what Justin says. Citing a discussion of the religious 
value of honey by Numenius and Cronius, Porphyry invokes the role 
played by honey, a pure and incorruptible substance, in the initiations 
to the Mithraic grade of Leo: 

otav nev oi>v xoic, ra A,eovnKa iiwuiaivou; ei<; xaq xetpa<; dv0' i)8aTO<; 
|ieA,t viyaaSai eyxscocn, Ka0apa<; e^eiv xaq X £ iP a ? rcapayye^Aovjcnv anb 
7tavTO<; AAmripcfu Kai P^omtKcn) kou iruaapo'u . . . KaGaipoucn 8e kou xr\v 
yXaaaav %$> iiiXui anb navxbq a|aapT(oA,cn3. 148 

De antro 15 

Numenius' conventional distinction between hands and tongue, between 
action and words, evidently stands for the totality of a person's social 
existence. We may here emphasise just three points. The words used to 
describe moral iniquities are common not merely in Greek moral texts 
of imperial date, such as Epictetus, but also occur repeatedly in early 
Christian contexts: Mithraism seemed to have rejected the same types 
of asocial, self-seeking behaviour as did popular philosophy and Chris- 
tianity, attempting, like them, to model a sense of personal integrity, 
to clear a space for the individual seeking to reconcile the demands of 
maintaining social status with those of conscience. Secondly, the nice 
discrimination between different types of sin (the distinctions between 
to Jumripov, to (3A,a7iTiK6v, to rroocxpov and to a^apTcoXov at this period 
are by no means easy for us to grasp) suggests a considerable refine- 
ment of moral discussion within Mithraism and a certain interest in a 
kind of moral casuistry. Thirdly, purification is only necessary if one 
has polluted oneself by contact with something unclean in the world 
outside, in this case perhaps with women, indeed with anything that 
the world constructed by Mithraism wished to exclude. The mention 
of the mouth suggests that lying too was one of the failings scouted in 
the cult: in any hierarchical group, lying disrupts the mechanisms of 



148 "When they pour honey instead of water onto the hands of those being initi- 
ated into the grade Leo to purify them, they bid them keep their hands pure from 
everything that is wicked, harmful and defiling . . . They also purify their tongues from 
everything sinful" (tr. Arethusa ed.). 



196 CHAPTER THREE 

control and provides the subaltern with a means of escape from pres- 
sures to conform. We may surely conclude that Numenius' ritual, at the 
point of initiation into the main or 'normal' grade in the cult, indicates 
that the boundary in Mithraism between 'insiders' and 'outsiders' was 
expressed very largely in moral terms: initiation certainly included 
other types of ceremonies, but the sheer strangeness of 'washing' one's 
hands in honey, of all things, must have made this transition singularly 
memorable. For not only did it mark one's introduction into a world 
in which nothing could be taken for granted, where honey was not a 
saccharine but a form of fiery water, appropriate to one's new nature 
as a Lion, a creature of fire, but one where, at least ideally, only the 
moral demands of Mithras, now matter. 149 

In this context, a recent find is of great interest. In 1998 a small 
fourth- century mithraeum, partly cut out of soft rock, was discovered 
in fairly dramatic circumstances beneath a Christian church at a place 
called Hawarte some fifteen km north of Apamea in Syria. Between 
c. AD 360-380 its walls and ceiling were re-decorated with a series of 
important frescoes. 150 Three are particularly interesting in the present 
context, despite the fairly extensive damage they have suffered. That 
to the left of the cult-niche on the north wall shows a city-wall made 
of opus quadratum with an arch in the centre.' 31 On top of the wall is a 
series of six horrifying severed heads. In some cases the skin is pale, in 
others brown or dark. Their hair is long and tangled, their mouths hang 
open, their eyes are lifeless. From each head a yellow line protrudes at an 
angle, perhaps a shaft of light, perhaps a gilded lance, which confirms 



149 Note line 7 at S. Prisca: [qui] aur<ei>s humeris portavit more iuvencum, "who has duly 
borne the bull on his golden shoulders" (of Mithras): Vermaseren and Van Essen 1965, 
200 no. 7. Emphasis on more. 

150 Gawlikowski 1999; 2000; 2001; 2001 [2004]; the fullest published account is idem 
2007, which includes colour photos. A monograph is in preparation. It has proved 
impossible to obtain photos of the most interesting frescoes; the reader is referred to 
Gawlikowski 2007, 343-6 esp. Colour figs. 12, 9 and 10 resp. 

151 The image may be a simplified form of the 'camp-gate' reverse type, so common 
on Tetrarchic silver coinage and struck inter alia at Antioch and Alexandria: RIC 6 
(1973) 616-18 nos. 34-43 (Antioch); 661f. nos. 9-13; cf. EA. Dumser, The Aeternae 
Memoriae Coinage of Maxentius: an Issue of Symbolic Intent, in L. Haselberger and 
J. Humphrey (eds.), Imaging Ancient Rome: Documentation — Visualization — Imagination. JRA 
Suppl. 61 (Portsmouth RI 2006) 107-18. Whereas in the western types there is a door, 
it is typical of these eastern ones that there is simply an open archway, as at Hawarte. 
The reverse legend on these coins is usually virtus militum. 



SYSTEMS OF VALUE 197 

the fact that all have been mastered and overcome. 152 A seventh head 
has fallen from the wall and lies on the ground. Nothing like this has 
ever been found in a mithraeum before, so that contextualisation is dif- 
ficult and necessarily tentative. Gawlikowski himself has suggested that 
the scene represents an assault by Mithras on the gates of hell, or the 
defeat of a City of Darkness or Evil by the forces of Good, perhaps in 
some eschatological context of which we know nothing 1 ' 3 It has also 
been suggested that it recalls the descriptions of the horrifying deaths 
of the contract-breakers, both wicked spirits and human beings, at the 
hands of Mithras' innumerable weapons in the MiQra-Tast. Knowledge 
of this authentically Iranian tradition may have reached Syria from 
Sasanid Iran via the traditional points of contact between the Iranian 
and Graeco-Roman worlds such as Nisibis or Zeugma — we are after 
all talking about the period immediately following Julian's inglorious 
Persian expedition, a period of relatively intense cultural contacts (Gor- 
don 2001 [2004] 106-09). Although Mithraists in Syria would hardly 
have recognised any of what they may thus have learned about Iranian 
Mithra, they may have been able to respond to this theme, because 
it resonated with an important feature of the Roman cult, the battle 
against the unjust, the evil-doers. On this view, the scene would have 
no eschatological significance but simply be a form of celebration or 
praise of Mithras' saving intervention against evil (which is anyway one 
of the connotations of the epithet dviicr|xo(;, invictus). 154 

At Hawarte, the entrance-door to the mithraeum was not, as usual, 
opposite the cult-niche but set into the west (left-hand side) wall. To the 
left and right of this entrance-doorway immediately visible to anyone 
wanting to enter the mithraeum, there was a symmetrical fresco, now 
known as The Twin Riders (Gawlikowski 2007, 353 no. 2). In each 
case, only the lower part is preserved; that on the left of the doorway 
is more complete than the one on the right. A person dressed in 
magnificently embroidered, aristocratic Parthian tunic and trousers 
is standing facing the spectator in front of a splendidly-caparisoned 



152 Gawlikowski has reiterated his view of that they must be shafts of light (2007, 
355); but if so it is strange that they are all the same length and are not shown as 
extending to the upper frame. 

153 1999, 203f; 2001 [2004] 187f; 2007, 355. 

1,1 That is not to say that an eschatological reference can be entirely excluded: 
Mithraic eschatology was undoubtedly more complex than many allow, and different 
fates may have awaited one depending on the standing of the qualities one acquired 
or refined in the eyes of Mithras. 



1 98 CHAPTER THREE 

white stallion. To the right is the base of a tall bronze candelabrum 
or turibulum encircled by a snake. Gawlikowski suggests this must be 
a double image of Mithras (2007, 353); alternatively, they might be 
protective angels or minor divinities, though certainly not Cautes and 
Cautopates. The two paintings are not however exactly symmetrical. 
The one on the left is holding a two-headed, or Janus-faced, black 
man, who is crouching or dancing, and kept firmly shackled by means 
of a chain attached to each wrist. The other end of the chain is held 
by the Rider, who also holds a stick of the kind one might use to keep 
herded animals in order. 153 

The quality of these paintings, and the interest in aristocratic dress 
and fine horses, suggests something of the likely social status of the 
owners of the mithraeum (assuming that these frescoes in the pronaos 
are indeed connected with the temple). This is definitely not a small- 
town or humble milieu. The image of a black demon kept in chains 
had by this date been familiar in folklore in Asia Minor for at least a 
century and a half. In the Acts of Peter 22, written in the late second 
century, the senator Marcellus dreams that a horribly ugly, pitch-black 
('Ethiopian') female demon danced in front of him, with chains round 
her neck and attached to her wrists. Although the significance of the 
Janus-head is unclear, the blackness of the demon at Hawarte is clearly 
negative. 1 ' 6 The function of the Riders must have been to protect 
those entering and leaving the mithraeum from evil. The late date of 
the temple means that we have to treat this evidence cautiously, but it 
seems to suggest a situation similar to what we can glimpse in Julian. 
The images convey a sense of what the cult was supposed at this date 
to offer, which complements rather than contradicts evidence from ear- 
lier periods. Just as Julian's vision is not totally invented, so we should 
not dismiss Hawarte on the grounds that it derives from a late phase 
when the cult was more or less moribund. It is rather the expression 
of a belief that this world is suffused with evil against which Mithras 
and his allies fight, and where his worshippers can trust implicitly in 
his victory 1 " 



155 Gawlikowski 2007, 353 with colour fig. 9 on p. 343. 

156 Lucian, Philopseud. 16 and 32; Epist. Barnab. 4.9; 20.1, cf. Gordon 2001 [2004] 
11-14. 

157 The mythical image of Mithras carrying the bull on his shoulders, the Transitus 
i, has thus apparently been interpreted in moral terms: Gordon 1988, 61. 



SYSTEMS OF VALUE 199 

The existence of an initiatory grade, albeit a low one, named axpoc- 
xid>xr\q/ miles may help to confirm the argument. The preparatory rituals 
for this presumably prepared the candidate physically and morally for 
the task of overcoming evil. As Turcan has pointed out (2000, 82) the 
metaphor of militia in a moral sense was widespread among philosophers 
and mystery-religions in the Principate; but Mithraism seems to have 
laid special emphasis on it. 1 " 8 Such rituals increased the discipline and 
solidarity of the group, the sense of serving a common goal, but also 
the idea of life as a battle. 159 

The third interesting fresco at Hawarte, likewise heavily damaged, is 
on the north wall of the pronaos (i.e. to the left of the entrance to the 
mithraeum), above the dado. It represents two beautifully-painted over- 
sized lions facing one another; each is attacking and killing a black man. 
The best preserved of the latter, naked except for a red breech-clout, is 
falling backwards as the lion leaps on him; blood drips from his neck; 
the head of the victim of the left-hand lion is severed from the bleed- 
ing trunk. 160 The iconography of such a scene is evidently borrowed 
from the amphitheatre (the lions are apparently being supervised by a 
figure in the background, who may be a tamer or trainer), but it again 
has interesting implications for our present theme. In Gawlikowski's 
view, the black men must represent evil demons being subdued and 
destroyed by the forces of light. 161 Given the twin Riders and the ghastly 
heads, this would constitute a consistent programme thematising the 
struggle between good and evil. By implication the lions would be the 
servants of Mithras, as evidently on the well-known hunting-scenes in 
the mithraeum at Dura-Europos, where a large lion is running directly 



158 Tertullian's account of the supposed ritual of the rejection of the crown, how- 
ever, simply implies that the Mithraic Miles was expected to declare his preference for 
Mithras over something which is not specified {De corona 15. 3f. = Sanzi 2003, 414: 
Mithras no. 5.3). There is no reason to suppose that the sword plays an important part 
in the scenario; it is just one of several oddities about the passage. 

139 "Leur apprentissage de l'endurance physique au feu et au froid n'en fait non 
plus des ascetes desincarnes et coupes de l'humanite commune: tout au contraire!": 
Turcan 2000, 1 13; Alvar 1993a, 129. The Miles who leads the procession on one face 
of the Mainz SchlangengefaB is dressed in full 'Hellenistic' armour, though without a 
helmet (Text-fig. 4, p. 347). 

i6o 'pjjg f resco i s very badly preserved; the clearest account is now Gawlikowski 2007, 
354 no. 3 with colour fig. 9 on p. 343. 

161 Gawlikowski 2007, 354; earlier (2001 [2004] 188) he was inclined to see it as 
representing an eschatological battle between Light (= the Lions) and evil. 



200 CHAPTER THREE 

in front of Mithras' galloping horse. 162 Nevertheless the Hawarte scene 
goes much farther, transforming these helping lions into dramatic images 
of the destruction of evil. 

By implication, victory over evil is not something haphazard, but, 
like the fortuna bestowed by Isis, a result of discipline achieved through 
pietas. Violent efforts are needed to succeed, but they too must have their 
proper measure. The lines from S. Prisca that refer to Mithras carrying 
the bull, and the reference there to the maxima divum (pp. 154, 375), 
present him as a moderate, disciplined hero, who carries out the duties 
assigned him by the gods. From this point of view, the tauroctony can 
be understood with reference to its iconographic model, Nike/Victoria 
sacrificing a bull. Here pietas and victoria become the supreme values 
represented in the paradigmatic ritual (cf. Zwirn 1989, 15). The Mithra- 
ist discovers in the cult-image both the core ethical message and the 
ideal model for his own behaviour. 

It is moreover likely that specific grades, where they existed, were 
associated with, or in some way particularly responsible for, specific 
virtues. Mithraic ethics were apprehended in two complementary ways. 
Peer pressure, expressed above all in the synecdoche of the hand-shake, 
provided a stimulus to maintain the required moral demands. At the 
same time there seems to have been a process of ostensive instruction, 
conveyed in ritual. On the basis of the evidence provided, mainly at 
Dura-Europos, by the moral adjectives and images associated with each 
of the grades, M. Meslin has set out what he sees as the connections 
between the believer's moral progress and his advancement up the 
ladder. 163 This is probably too schematic, however; individual epithets 
were in all probability quite localised in their use, another example of 
the tendency in the cult towards 'idiolectal' coinages. 164 Given that it is 
practically all we have, however, the evidence from Dura, where there 



162 Rostovtzeff 1939, 1 12f. with pi. XIV 1-2. In the fresco on the North wall there 
are in fact two helping lions, since a small lion has replaced the helping snake that 
figures in the version on the South wall. No lion (and no Mithras) is preserved in the 
hunting scene on the South wall of the Hawarte mithraeum (Gawlikowski 2007, 358 
no. 18 with figs. 17 and 17a on p. 359). 

163 Meslin 1985, 179ff; cf. Gordon 1972, 106-10; Francis 1975, 441-43; also the 
table of the epithets at Dura, Gordon 1980, 41. 

164 Cf. R.L. Gordon, Mystery, Metaphor and Doctrine in the Mysteries of Mithras, 
in Hinnells 1994, 103-24. 



SYSTEMS OF VALUE 201 

clearly was a system of grades, even if there may not have been seven, 
needs to be given due weight. 165 

The individual initiate {syndexios) is regularly given the epithet ayoc- 
96c;. lf ' 1 ' Though hardly a discriminating word, this makes the general 
moral aim of Mithraism quite clear. 1 '" At Rome, by contrast, the indi- 
viduals composing his Mithraic group as a whole are described by the 
founder of one late third-century mithraeum as hilares, joyful. 168 We 
might speculate that the 'message' to be spread by the Corax, Raven, the 
lowest grade, named for the raven that brought Mithras his message, 
perhaps concerned the virtues of Mithras and his cult. 169 The grade 
Miles surely applied the military virtues to the Mithraic brotherhood; 
that in turn must have encouraged an elective affinity between the cult 
and the army. At Dura, the grade's usual epithet is dKepaioq, perfect, 
uncontaminated, pure. 170 It must be admitted, however, that from this 
point on the epithets become less concerned with moral qualities, more 
with the internal logic of the grade. Thus the Duran epithet for Leo is 
6c(3p6q, splendid, fresh, charming 171 Porphyry tells us that Perses, the fifth 
grade, for which Dura provides no usable evidence, was connected with 
the protection of the harvest (<pi>Xa£, xapncbv) and was therefore also 
connected with honey, because it keeps things from going rotten. 1 ' 2 At 
Dura there is no grade Heliodromus, and oxepecoxriCj, whatever it means, 
seems to do duty for it; the standing epithet seems to be acxpioxric;, 
which certainly implies a virtue, but hardly a moral one. 173 



165 On the problems raised by the evidence for the grades, see Chap. 4.5.b below. 

166 Rostovtzeff et al. 1939, 120 no. 858 = V. 60 (c. 250 AD). Another very general 
epithet is zx>oe^r\c,: Rostovtzeff 1939, 124; again it can hardly be considered anything 
more than a vague pro-word. 

"" Meslin's claim that the word occurs at Dura in relation to Nymphus, the second 
grade, is a mistake due to taking over Vermaseren's entry s.v. V 63, without checking 
Rostovtzeff 1939, 124, where Cumont does not mention the matter, and indeed approves 
Wickstead's suggestion that the ordinary epithet at Dura for Nymphus is veoi;. 

168 AE 1937: 231 = 1950: 199 = V 423, syndexi hilares. A relationship between this 
state of mind and initiation is clearly implied by the rest of the text, a state that evi- 
dently distinguishes initiates from non-initiates. 

169 Another mid-IIP votive, from the Emporium area of Rome, describes a Corax 
as iepoi;: IGVR 107f = V 473f. This epithet later became stereotyped: for example, the 
late fourth-century Palazzo Marignoli texts speak of initiations named hierocoracica (CIL 
VI 751b = V. 403 (376 AD). This was evidently a development intended to distinguish 
the grade from the bird. iep6<; apparendy does not occur at Dura. 

170 Rostovtzeff et al. 1939, 120 no. 857 = V 59. 

171 Rostovtzeff et al. 1939, 120 no. 856 with p. 124 = V 58. 

172 De antra 16 = Sanzi 2003, 420: Mithras no. 8.4; Merkelbach 1984: 115f 

173 Rostovtzeff et al. 1939, 123. 



202 CHAPTER THREE 

It thus seems that emphasis was laid in the early stages of member- 
ship of a Mithraic group on fairly anodyne, generalised and non-specific 
moral qualities, which nevertheless pointed in a consistent direction. It 
goes without saying that such qualities were also in keeping with the 
dominant official ethics of the wider society Of greater de facto impor- 
tance than these rather formulaic virtues, I think, were the qualities tac- 
itly encouraged and favoured: obedience, self-discipline, self-examination 
and moral seriousness, in order to gain the attention and appreciation 
of the god. The cult thus tended to favour social risers. 174 

However in a complex society desires and aims are extremely var- 
ied — even antithetical to one another. They cannot be meaningfully 
reduced to a single ethical scenario. As I stressed at the beginning of 
this chapter, a variety of value-systems may co-exist, not merely in the 
wider social formation but also in any one cult. The real situation was 
undoubtedly infinitely more complicated than what we can ascertain on 
the basis of our very limited information. Despite its notorious gender 
exclusivity, however, the integrative potential of Mithraism is clear. It 
was not troubled by the type of sexual or marital issues that we have 
seen causing problems in the other cults, nor was it denounced for 
undermining the family or seducing the young from the path of virtue, 
as happened in the case of Christianity. 1/S Indeed, part of the cult's 
appeal was its masculine exclusiveness, which favoured the development 
of a specifically male ethic of asceticism, an ethic fitted for those for 
whom the sentimentalisation of marriage held no attraction. 

Just two further points here. The first concerns the cult's aims. Many 
people erroneously suppose that all religions have a sort of universalist 
tendency or ambition. In the case of Mithraism, such an ambition has 
often been taken for granted and linked to another no less questionable 
assumption, that there was a rivalry between Mithras and Christ for 
imperial favour. Despite Diocletian and the Tetrarchs' famous dedication 
D(eo) S(oli) Ifnvicto) M(ithrae) fautori imperii sui at Carnuntum in AD 308, 
it is far from clear that there was ever any desire on the part of the 
political centre to turn Mithraism into the main ideological support of 



174 On all this, see Gordon 1972; Merkelbach 1984, 160f.; Meslin 1985, 179-81; 
Clauss 2000, 39-41. 

175 This was one of Celsus' criticisms mentioned by Origen, Contra Cels. 3.55; it 
tells us something about the methods of proselytism, cf. J.A. Francis, Subversive Virtue: 
Asceticism and Authority in the Second-Century Pagan World (University Park PA 1995) 156f.; 
also the essay by my late friend F. Gasco, El Pastor de Hermas y la familia, ARTS 1 
(1998)207-11. 



SYSTEMS OF VALUE 203 

the regime. 176 The various strands of the public cult of Sol Invictus had 
little connection with the homonymous divinity in Mithraism (Berrens 
2004). If Christianity had failed, the Roman Empire would never have 
become Mithraist. 1 ' 7 This doubt about the cult's universalist aims leads 
on to the second point. The reverse of the appeal to a certain kind 
of man was that the cult sharply restricted its potential audience and 
could therefore hardly develop into a genuinely universal religion. The 
cult's failure to recover after the mid third century, despite some efforts 
by imperial officials to re-establish it, not least in the army, must have 
been due largely to its self-imposed gender restrictions." 8 



176 CIL III 4413 = ILS 367 = V. 1698. 

'" I refer of course to the famous flourish by E. Renan: "On peut dire que si le 
christianisme eut ete arrete dans sa croissance par quelque maladie mortelle, le monde 
eut ete mithriaste", Histoire des origines du christianisme, VII: Marc-Aurele et la fin du monde 
antique (Paris 1883) 579. 

178 For the efforts to restore the cult after mid-century, see Clauss 2000, 28f. 



CHAPTER FOUR 
THE RITUAL SYSTEMS 

El cuerpo ha creado el espacio, como el agua crea 
el recipiente. 

Sherezade 1 

We come now to the sub-system in my original scheme for which we 
possess the most extensive evidence and which has therefore received 
most attention in the scholarly literature. After some introductory 
remarks on the function of ritual in general, I propose systematically 
to discuss the main rituals practised by each of the mysteries. It may 
however be as well first to say something about the notion of ritual 
within the wider context of a religious system. Ritual has been defined 
as a "type of behaviour characteristic of, and recognised by, a given 
group, whose repeated performance by individuals is taken as a sign 
of their membership". 2 



1 "Solid bodies gave rise to space just as water gives rise to its receptacle". Quoted 
from the translation of 1001 Nights by P. Martinez Montavez (Madrid 1977) 108. 

2 E Marty, Le rite et la parole, in AA.VV 1981b, 67. I do not think this an ideal 
formulation but it is sufficiently elementary not to arouse too many hackles. It has the 
merit of raising the issue of behaviour, which is always historically conditioned (albeit 
possibly over the longue duree), rather than appealing to timeless continuity. I would 
however be inclined to replace the word behaviour with the phrase "act or acts" or 
even "gestures or series of gestures", since these terms highlight the importance of 
detail in ritual analysis. Marty's formulation also picks up the social element of ritual, 
even though it may be performed by a single individual as well as by a collectivity. 
Finally, it refers to the function of ritual, its role in maintaining social cohesion. There 
is no need to stress 'religious' here, since the members of a football team or a group 
of pop fans may develop their own rituals. However, the functions of ritual are in fact 
very varied, so that different perspectives will legitimately highlight different accounts. If 
we take prayer, for example, it is hard to decide whether its main function is to affirm 
group cohesion or to set up a specific relation between the person who performs the 
prayer and the supernatural. One conception of 'function' will stress the first, another 
the second. If we take all the individual cases together we might arrive at yet a third 
view, that its function is not to maintain cohesion but difference, since individual desire 
can more easily be manipulated in a prayer-regime. If this seems excessively cynical, 
think of Q. Mucius Scaevola (cos. 95 BC = RE Mucius no. 22), who, though pontifex 
maximus, is said to have claimed that, of the three types of deities, those of the poets, 
the philosophers and the principes civitatis, only the last were of any value, on the grounds 
that it was necessary that the ordinary people should be deceived in matters of religion, 
expedire . . .falli in religione civitates (Augustine, De civ. Dei 4.27 = IAH 1 p. 102f no. 7 = 



206 CHAPTER FOUR 

In addition to sociological definitions of this kind, the role of emotions 
and shared experience in the ritual process has recently been stressed. 3 
What interests me about the rituals of the oriental cults is the evidence 
they offer of the dialectical relation between two opposed tendencies, 
towards social integration on the one hand, and towards continuing 
marginality on the other. I see these tendencies working both at the 
individual and the collective level. As regards the individual, personal 
transformation had to navigate between liminality and participation, 
both during one's own initiation(s) and in continuing membership of a 
cult whose politico-religious status was ambivalent. At the group level, 
the mysteries were historically conditioned by the ambivalent attitudes 
of the wider society; their social catchment meant that they could never 
enjoy an unequivocal status. Up to a point, we can summarise this 
institutionalised ambivalence by invoking Victor Turner's contrasting 
notions of liminality and communitas.* 

1 . Religion and Ritual 

Aiyco 8e 7t?uxc>|j.aTC>8e<; to 7tpo<; vnoQeaw 
(3e(3ida|ievov 

Aristotle, Met. M 7, 1082b3f. 5 



Varro RD 1 frg. 7 Agahd — frg. 27 Condemi, relegated by B. Cardauns to the appendix 
to the fragments of Bk 1 [1 p. 37 frg. V; cf 1 p. 18 under frg. 7; cf. 2, pp. 139-43] on 
the grounds that it may be from Varro's Logistoricus Curio de cultu deorum). 

3 Chaniotis 2006, 211-38; S. Scullion, Festivals, in Ogden 2007, 190-203 at 
201-3. 

1 Cf. VW. Turner, The Ritual Process. Structure and Anti-Structure (London 1969). Lim- 
inality, derived of course from limen, corresponds to the notion of marginality, not in 
the sense of something foreign but as the boundary or border structurally required to 
define a centre, communitas in Turner's jargon, the nucleus or core of the system. Turner 
not only defines both concepts but illustrates the inter-relation of the liminal with the 
communitas by means of concrete examples. I do find the general scheme stimulating 
but it is difficult to apply his work directly to my case, since the rituals he deals with 
are a mainstream part of the cultures he is discussing, with their own specific religious 
reality: analysis of the rituals aids in understanding the wider society. In my case I 
am dealing with adaptations of foreign systems inserted secondarily into a pre-existing 
socio-cultural order, so that there can be no question of a precise fit between the two. 
I also rather share the doubts of Bourdieu 2001, 175 about whether Turner has in 
fact added much of enduring value to van Gennep's insights. 

5 "By a fiction I mean a forced statement made to suit a hypothesis" (tr. W.D. 
Ross). 



THE RITUAL SYSTEMS 207 

This is not the place for an extensive enquiry into Aristotle's claim; I 
want simply to make a few points relating to the place of ritual in a 
given religious system and a given social context. 6 Myth cannot exist 
without a society to tell it. The performance of ritual by a group or its 
constituent individuals will be still more a matter of deliberate intent, 
since ritual is directly linked to the everyday life of the individual, family, 
institution or whole society. Correct performance is the sole guarantee 
that the ritual will not become distorted: wilful alteration is inhibited by 
the requirement of fidelity to a specific sequence of actions. At the same 
time, individual participation, personal performance, provides the crucial 
link between ritual and religious thought. 7 Rituals supply the stage where 
the drama of the myth is played out. 8 They are the space, as Loisy put 
it, where humans collaborate actively in the gods' work. 9 

On the other hand, the ritualisation of religious ideas is a highly 
effective means of ideological control. Virtually all types of claim to 



6 Bell 1997 is an excellent introduction, mainly for students, with a clear overview 
of the development of scholarship in this area, and of modern trends. I have also 
made use of many other discussions. Among these I may mention: F. Bousquet, "Et la 
chaire se fit Logos...". Essai sur la reaction philosophique au rite, in AA.VV 1981b, 
33—66. Despite its promising title, I would not recommend O. Wikstrom, Liturgy as 
Experience. The Psychology of Worship: A Theoretical and Empirical Lacuna, in 
Ahlback 1993, 83-100. 

' The relation between individual and collectivity in ritual contexts has traditionally, 
since Freud's Das Unbehagen in der Kultur (1930), been the focus of psychology. The work 
of R. Girard, particularly Violence and the Sacred (Baltimore and London 1977, orig. 
ed. 1972), can be seen as a direct continuation of this line of thought; cf. J. Ries, Le 
sacre et l'histoire des religions, in idem (ed.), L'expression du sacre dans les grandes religions, 
1: Proche-Orient ancien et traditions bibliques. Homo Religosus 1 (Louvain-la-Neuve 1986); 
J. Greisch, Une anthropologic fondamentale du rite: Rene Girard, in AA.W 1981b, 
89—119; F. Dumas-Champion, Le sacrifice ou la question du Meurtre, Anthropos 82 
(1987) 135-49; there has even been conferences devoted, at least partly, to his ideas 
and methods: P. Demouchel (dir.), Violence et verite autour de R. Girard. Colloque de Cerisy 
(Paris 1985); R.G. Hamerton-Kelly (ed.), Violent Origins. Ritual Killing and Cultural Formation 
(Stanford 1987). Some of his sympathisers find applications of his ideas in the myths 
of Cybele, Isis or Mithras; but from my point of view they are quite unusable. The 
idea of a mimetic reproduction of aboriginal violence makes some sense of individual 
behaviour, I suppose, but not of collective action, which needs to be explained in quite 
different terms. This is of course also the problem with the central theses of W. Burkert, 
Homo Mecans:The Anthropology of Ancient Greek Sacrificial Ritual and Myth (Berkeley 1983) 
[orig. ed. 1972], since the means, and point, of the transmission from the Palaeolithic 
remain wholly obscure. For recent and current tendencies, cf. Rudhardt and Reverdin 
1981; Grottanelli and Parise 1988; H.S. Versnel, Ritual Dynamics: The Contribution 
of Analogy, Simile and Free Association, in Stavrianopoulou 2006, 3 1 7—28. 

8 Cf. Sami-Ali, L'Espace imaginaire (Paris 1974) 15f; at greater length: G. Durand, 
Les structures anthropologiques de I'imaginaire (Paris 1992). 

9 A. Loisy, Essai historique sur le sacrifice (Paris 1920) 61. 



208 CHAPTER FOUR 

hegemonic authority involve such ritualisation. The religious behaviour 
of the masses is thereby subjected to the control of those who claim the 
right to manage relations between men and gods. Institutionalised reli- 
gion thus has a key role in the formation and consolidation of relations 
of inequality established by state systems. Where religious experience is 
not subject to such regulation, we may properly speak of religiosity. In 
other words, religion is found in societies where there is a state, whereas 
non-state societies see no need to control relations between individual 
and the gods so tightly. Religion thus involves the control of religious 
impulses, which of course implies nothing about the complexity of 
the latter: as we know from the famous cases of the Nuer, the Zande, 
the Dinka and the Dogon, the religiosity of non-state societies may 
be at least as complex as the religion of stratified societies. Since for 
me religion can claim no higher ethical or other value, but is simply 
a secondary product of the drive towards institutionalised inequality, 
there can be no question here of asserting an evolutionary scheme 
from savagery to civilisation. Nor do I view the relation between the 
universe of religious thought and its individual or collective manifesta- 
tion in terms of qualitative, i.e. value-laden, differences. 

Once religion has established control over the expression of reli- 
gious feelings, rituals end up by regulating more or less the whole 
of every-day life, so that every social act (e.g. birth, marriage, death, 
agreements, performatives), and every aspect of production (e.g. sow- 
ing, harvesting, making pots, sailing) has to be accompanied by some 
appropriate sacralising performance. The rituals come to dominate 
human life. Conversely, structural changes in society demand changes 
at the ritual level, whether by the introduction of new forms, or by 
attributing new meanings to existing rituals. The ritual order is thus 
by no means static but subject to the demands of changing social 
needs. 10 This means that the analysis of ritual is extremely tricky. The 



10 Cf. Turcan 1992b, 215: "Ces mysteres ne se perdent pas dans la nuit des temps 
et ne precedent pas (ou pas settlement) d'ttne reamenagement des liturgies liees aux 
initiations 'primitives' mais . . . il est arrive un moment oil des hommes et les femmes 
ont fabrique un rituel pour repondre a un besoin, aux mutations des mentalites ou de 
conditions de vie". J. Scheid, Rituel et ecriture a Rome, in A. Blondeau and K. Schip- 
per (eds.), Essais sur le rituel, 2: Colloque du Centenaire 1987 (Paris and Louvain-la-Neuve 
[1995]) 1-15, argues that there were no full transcriptions of rituals at Rome, despite 
the existence of well-known texts on these topics, since the details were transmitted 
by word of mouth; as Numa claimed: ot> koX&q, ev d\|r6x 01 ? Ypd|J.|J.aai <ppov)pot)|xev(ov 
twv a7toppr|i:(ov, it is not right that such religious arcana should be kept in lifeless texts 
(Plutarch, Numa 22.2). G. Dumezil, La tradition druidique et l'ecriture: le vivant et le 



THE RITUAL SYSTEMS 209 

ritual may appear to contain a sacralised version of ancient elements 
of the myth (whether with reference to the time of composition or to 
the imaginary Mud tempus) but in fact offer something new, either a re- 
reading of the symbolism amounting to a new interpretation, or the 
adoption of something completely novel. 11 

For Durkheim, ritual serves to suffuse the profane with the sacred. 
The idea of society is the 'soul of religion'. The sacred is a projection 
of the collectivity just as the ritual process is the sacralisation of the 
community's own actions. But this is to put the cart before the horse: 
such a formulation absurdly claims that there must have been a collective 
agreement to sacralise the community. So far from valorising relations 
of social solidarity, the Durkheimian view subordinates everything to 
the community, so that the established order is hypostasised beyond all 
human action and thought; the individual plays only a negative and 
passive role. Religious life rests "at least partly on the failure to recognise 



moit, in Cahiers pour un temps (Paris 1981) 325—38: "C'est done par les yeux et l'oreille 
que le neophyte s'initiait aux gestes rituels precis, de sa jeunesse, aux cotes de son pere". 
On the other hand, it is familiar that in the Hellenistic period at least some attempt 
was made by kings, notably the Ptolemies already in the third century BC, to control 
private (Dionysiac) religious groups by requiring the presentation to the authorities 
of a sealed (if the restoration is correct) copy of the hieros logos and a statement of its 
unchanged transmission through three generations: BGU 1211 — SB 7266, tr. Hunt 
and Edgar, Select Papyri (Loeb) 2 no. 208, with M.-T. Lenger, Corpus des ordonnances des 
Ptolemies. Memoires de l'Academie royale de Beige 57 (Brussels 1964) 68f no. 29; Fraser 
1972, 1: 204 with nn. 114f. Such requirements may be the origin of written texts in 
these cults during the Imperial period, when it seems clear that such knowledge was 
available, either by profane means, such as novels, or through liturgical texts, in the case 
of the cult of Isis (cf. Clement of Alexandria, Strom. 6.4.35), and possibly Mithras. 

11 There are two main dynamic approaches to ritual. On one view, its eifect on 
the external world is to alter actual practice, as suggested in the main text; alterna- 
tively, it may be understood as a sort of time-regulator, such that the dynamics of the 
ritual have a knock-on effect in the wider culture (all this is quite independent of the 
famous debate over cyclical versus linear conceptions of time started by M. Eliade, 
The Myth of the Eternal Return, or Cosmos and History [New York and Henley-on-Thames 
c. 1954; orig. ed. 1948], on which see AA.W, L'eternel retour: conferences et debuts. Publ. 
de la Faculte des Lettres, Arts et Sciences Humaines de Nice, no. 9 [Paris 1992]). See 
D. Sabbatucci, Storia e metastoria, in idem 1978, 15-32; P.-J. Labarriere, Le rite et le 
temps, in AA.W 1981b, 13-32; E. Campi, Rite et maitrise du temps, in Centilivres 
and Hinard 1986, 131-37: "Nous partons du postulat selon lequel la necessite de 
certains rituels va de pair avec la necessite de controler le temps . . . Temps et rituels 
se repetent ainsi de facon circulaire, dans un eternel retour a un prototype mythique 
qui evacue toute reference a l'avenir, un avenir impensable puisqu'il ne se differencie 
pas de ce qui a ete. Or nier l'avenir c'est abolir le temps et c'est surtout nier la mort" 
(p. 13 If). I would particularly recommend G. Mazzoleni, II tempo ciclico riconsiderato, 
in Bianchi 1986b, 69-77. 



210 CHAPTER FOUR 

that religious ideas are symbolic representations of social reality". 12 As 
many have pointed out, this is absurd. 13 I would argue that the sacred 
cannot be the projection of a social consensus but is rather the opposite: 
it is a theoretical construction to justify the on-going development of 
social inequality projected as the responsibility of beings of another, 
and superior, order of existence. There is thus no point in mere humans 
kicking against the pricks. The fiction, Aristotle's xb nXaa^ax«>8ec„ is 
taken as a straightforward description of a fact. However, as in the 
case of Rousseau's social contract, the beneficiaries of the construct 
are all to clearly involved in its formation. Once again, the question 
cui bono retains its value in the analysis. Conversely, ritual, as the cyclic 
reproduction of this symbolic order, helps reinforce perceived reality; 
and when the latter changes, as it is bound to do, retains its power to 
act as an ideological brake by interpreting change as a manifestation 
of divine will. 14 

In the light of such considerations, it is only from the perspective 
of the interests of the socially privileged that we can understand 
Cazeneuve's claim that religious thought presupposes the existence of 
a principle that sustains the human condition, a principle mediated by 
ritual. 10 For such a claim presupposes that all members of the 'human 
condition' live under the same objective conditions; alternatively, if 
we are to assume that all human beings have the same 'real' interests 
irrespective of their objective situation, we would have to invoke some 
artificial homogenising agent capable of making these interests similar. 
In itself, mere participation in ritual cannot ensure that the 'principle' 
sustains everyone equally, ritual being notoriously concerned with 
hierarchy and status. 

On the other hand, every religious system fosters different ways, 
individual and collective, of approaching the divine. Many different 
situations may co-exist within the same system, as we can see from 



12 Skorupski 1976, 35; 187. 

13 E.g. S. Lukes, Entile Durkheim. His Life and Work. A Historical and Critical Study (Har- 
mondsworth 1973) 477—84; of course, Durkheim's view of religion is anthropocentric, 
but only in the sense that it is not cosmocentric: Skorupski 1976, 3 If. 

" Etymologically, 'rite' and 'ritual' are derived from Latin ritus, religious observance(s), 
ceremonies, adj. ritualis, which in turn seems to share a common IE sense with Skr. flam, 
Avestan ahm, both meaning 'order, in conformity with what right practice demands': 
Ernout-Meillet p. 574 s.v; Walde-Hofmann 1: 437 s.v; cf. J. Ries, Rites d'initiation, 
in idem 1986b, 30. 

15 J. Cazeneuve, Sociologie du rite (Paris 1971) 282f 



THE RITUAL SYSTEMS 211 

the fact that the rituals of Greek and Roman civic religion are very 
different from those specific to the oriental cults. At the same time, as 
social complexity increases, relations between gods and humans acquire 
more complex forms of expression. That is why ritual can properly be 
examined from different theoretical perspectives in accordance with 
individual inclination. 16 My own view is that rituals are not independent 
or autonomous forms of action but only make sense within a larger 
symbolic framework. They are therefore best understood as perfor- 
mances or dramatisations whose function is to connect the real with the 
imaginary in a variety of ways, thus conferring life and vitality upon the 
fiction of the divine Other. It is the direct experience of such enacted 
performance that enables the believer to accept the fiction as though 
it were indeed from time to time part of his or her lived reality. 

2. Rituals in the Mysteries 

Nolite annuis sacris quaerere funus alienum. Vestris potius 
funeribus parate solaria per annos singulos. 17 

Firmicus Maternus, De errore 2.8 (ed. R. Turcan) 

The fact that everyday experience and the fiction of religion are con- 
stantly in dialectic with one another must inhibit any attempt to offer 
a monolithic account of the rituals of the mystery-component of the 
oriental cults. It would of course be much easier if one could, certainly 
at the level of interpretation, but the operation of the dialectic over 
time makes change inevitable. It is therefore methodologically unsound 
to assume that a ritual existed unchanged long before or long after the 
point at which we have information about it; but equally we cannot 
argue that the absence of a given ritual at one period means that it 
could not have existed at another — a point of particular importance 
when we come to rituals known only from inherently unreliable Chris- 
tian sources, which are never mere 'reports', and always have their 
own distorting agenda. On the other hand, our documentation is so 



"' Useful further discussion in L. Bouyer, Rite and Man: Natural Sacredness and Christian 
Liturgy, tr. J. Costelloe (Notre Dame 1963 — Le rite et I'homme: sacralite naturelle et liturgie 
[Paris 1962]); J. Greisch (ed.), Le rite. Philosophic 6: L'Institut catholique de Paris (Paris 
1981); Bell 1997. On psychological approaches, see B. Beit-Hallahmi and M. Argyle, 
The Psychology of Religious Behaviour. Belief and Experience (London and New York 1997). 

" "Give up seeking someone's else's corpse each year: you would do better to spend 
your time year by year seeking comfort in the face of your own death!" 



212 CHAPTER FOUR 

lacunate that there can equally be no question of attempting a proper 
history of ritual development, except in a few cases, say the taurobolium/ 
criobolium, where there is sufficient epigraphic evidence to reconstruct 
a historical development, though even here, as we shall see, sharply 
different accounts have been offered. A fundamental problem that we 
shall encounter repeatedly in the following pages is the legitimacy of 
the traditional habit of generalising about ritual practice, mystery and 
other, on the basis of highly fragmentary evidence. 

It used to be taken as an article of faith that official or civic religion 
in antiquity was largely distinct from the true feelings of the people. 
The idea was that the success of the mysteries was due to the fact 
that their rituals made a more direct appeal to their adherents. As I 
pointed out in the Introduction, this view needs substantial modifica- 
tion. MacMullen and Lane Fox showed in the 1980s that traditional 
public religion, both in the Latin West and in the Greek East, was by 
no means as alien from ordinary people as had often been imagined. It 
has been easy to show that city populations participated enthusiastically 
in public rituals throughout the Principate, just as they did earlier. 18 So 
much is by now commonplace. Nevertheless, even MacMullen was still 
inclined to distinguish between civic cult and 'dynamic cults', and it 
seems to me clear that official religion could not often offer experiences 
comparable to those available in the oriental cults, whose deities were 
able to communicate a particular sense of 'standing by' or 'being with'. 19 
Something of the dramatic impact of their rituals can be gauged from 
Firmicus Maternus' feeble advice in the epigraph to this section that 
pagans should think more about their own deaths than those of their 
gods. Inasmuch as rituals are a crucial means of constructing a rela- 



18 MacMullen 1981; Lane Fox 1986; Alvar, Blanquez and Wagner 1995; Beard, 
North, Price 1998, esp. chap. 7; B. Dignas, A Day in the Life of a Greek Sanctuary, 
in Ogden 2007, 165—77; N. Belayche, Religious Actors in Daily Life: Practices and 
Related Beliefs, in Riipke 2007d, 275—91. Changes certainly did take place here too, 
though the subtle work required to show them is not often done: K. Hopkins, From 
Violence to Blessing: Symbols and Rituals in Ancient Rome, in A. Molho, K. Raaflaub, 
and J. Emlen (eds.), Athens and Rome, Florence and Venice: City States in Classical Antiquity and 
Mediaeval Italy (Stuttgart 1991) 479-98. 

19 R. Turcan has however argued that the aims and modality of ancient festival largely 
overlapped with those of initiation: La fete dans les rituels initiatiques, in A. Motte and 
CM. Ternes (eds.), Dieux, fetes, sacre dans la Grece et la Rome antiques. Colloque de Luxembourg 
24-26 oct. 1999 (Turnhout 2003) 7-21. He highlights the cathartic effects of music, 
dance and song, collective eating and the busyness of the games. This seems to me not 
to make enough of the differences between the Eleusinian and later mysteries. 



THE RITUAL SYSTEMS 213 

don between men and gods, they will offer a spectrum of possibilities, 
choices and competences in this regard. 20 If we posit such a spectrum, 
mysteries of all types, but particularly those of the imperial period, 
will be well up towards the pole favouring the establishment of close 
personal relations between deity and adherent, indeed in a manner that 
would traditionally have been considered unmanly or even superstitious 
by the civic elite. 21 

This relation was however mediated by a process of learning. As 
one of the institutions characteristic of complex hierarchical societies, 
religion tries to avoid spontaneous contact between believer and deity, 
which is characteristic of the religiosity of non-state societies where 
intermediary roles are not so well developed. 22 The oriental cults, with 
their mysteries, developed in an historical context in which they could 
no longer refuse such regulated access to divinity, as is evident both in 
the specific composition of their rituals and in the formalised process 
of learning that enabled the initiand to become a full member of the 
group. As usual, the responsibility for this situation was attributed to 
the deity, since it is the gods who, through specific signs, such as those 
given in dreams, choose those who are considered worthy of being 
initiated (e.g. Apuleius, Met. 1 1.19, 21, 22, 26-7, 29-30). Thus begins a 
habit of submission all too accommodating to the interests of a highly 
fragmented state with rather few spare resources, beyond coin-types, for 
nurturing ideological commitment to itself. The hope of belonging to 
a select group, a 0(aoo<; of some kind, even if only for the duration of 
a ritual, which was to be met with again in the afterlife, was a further 
incentive to accept a degree of difference, even alienation, sweetened 
as it was by the promise of admission to a secret that only the chosen 
can share. 23 This presupposes a process of learning — or perhaps better 
'apprehension' in the sense 'a perception that is comparatively simple, 
direct and immediate and has as its object something considered to be 



2(1 Cf. V Pirenne-Delforge, Image des dieux et rituel dans le discours de Pausanias, 
MEFRA 116 (2004) 811-25. 

21 Veyne 1989, 185f.; 1999, 434-39; Henrichs 1998. 

22 Cf. the classic study by I.M. Lewis, Ecstatic Religion (Harmondsworth 1971 and 
often repr.). 

23 E.g. the 'new' funerary gold leaf from Pherai in Thessaly, where the dead 
(woman) requests the uncertain addressee 7:e)j,7te |ie Jtp6<; |Xi)cn:S[v] Oidootx;: Parker 
and XranciTOTTOuXot) 2004, 6, with their commentary p. 8: "It ... strongly suggests that 
the dead person may have been a member of a Giaooi;, at least for the duration of 
a particular ritual, and that the membership is part of his or her qualification for a 
blessed lot in the afterlife". 



214 CHAPTER FOUR 

directly and non-discursively understandable' 24 — sufficiently absorbing 
or attractive to transform objective oppression into subjective mystical 
delight. 

It is however extremely difficult to say quite how this knowledge 
or understanding was communicated. There is no direct evidence for 
formal instruction. Written hieroi logoi do seem to have existed in many 
mysteries but they are virtually all lost. 25 The only text that gives us a 
specific idea of what they may have looked like is the well-known PGurob 
1, an Orphic-Bacchic document of the third century BC, sadly consist- 
ing of just the ends of thirty lines from one column of an entire roll, 
and scraps of a second. 26 The surviving text presents a dense sequence 
of apparently brief prayers, sacrificial prescriptions, <x6|i|3oA,cx and ritual 
instructions that clearly presupposes first-hand knowledge of the actual 
procedures acquired through multiple exposure. It can in fact best be 
described as an aide-memoire for someone already very familiar with 
both ritual and sense. 27 In the cult of Isis, there were numerous types 
of written text, including bi- or even tri-lingual liturgies, for specific 
occasions: at least in fiction, it could plausibly be suggested that the 
liturgy at the ploiaphesia or Isidis navigium, consisting mainly of prayers 
for the Princeps, Senate, equites, the entire Roman people, and all mari- 
ners, was read word for word in Latin by the presiding priest, while 
the remainder of the ritual was conducted in Greek. 28 The S. Prisca 
texts and the scraps from Dura-Europus, as well as the three croupoAa 
cited by Firmicus Maternus, De errore 5.2; 19.1 and 20.1 (= Sanzi 2003, 
432—34, Mithras nos. 19.1—3) do imply the existence at least in some 
mithraea of similar hieroi logoi; the crucial difference however is that 
in Mithraism there were no public rituals, and therefore no need to 



24 The definition of Webster's Third New International Dictionary (Springfield 
MA 1961) s.v. 

25 "Books were used in mysteries from an early date": Burkert 1987, 70, distinguish- 
ing three historically consecutive types: myth, nature allegory and metaphysics. I have 
already noted the Ptolemaic ordinances (n. 10 above). 

26 See the commentary by J. Gilbart Smyly Greek Papyri from Gurob (Dublin and 
London 1921) 1-10 no. 1. The text is reproduced as Kern OF 101-04 frag.vet. 31 
(which also gives the scraps from col. ii, omitted by Smyly) = DK 1 F 23. It is per- 
fectly possible that PGurob 1 is an example of the kind of hieros logos required by the 
Ptolemaic ordinances. 

27 Cf "Script culture remains highly oral in its use of language and texts": W.A. 
Graham, Beyond the Written Word: Oral Aspects of Scripture in the History of Religion (Cam- 
bridge 1987) 17. 

28 Apuleius, Met. 11.17, cf. Finkelpearl 1998, 53. 



THE RITUAL SYSTEMS 215 

construct the appropriate ritual components. For the Phrygian cults, 
on the other hand, which had a plethora of public rituals, nothing 
similar survives, though Firmicus Maternus must surely have found the 
cruupoXoc that he cites from the cult in some such text. 29 

Hieroi logoi however tell us nothing about instruction. In one of the 
passages of Apuleius, Met. 1 1 that most clearly seems to support the 
modern ironic reading of the book, 30 the Isiac priest Mithras takes out 
from a safe in the temple some books in hieroglyphic characters whose 
fantastic shapes are designed to prevent prying by the ignorant, 31 and 
reads out to Lucius what he is to do to prepare himself for initiation 
(11.22). The reader is thus clearly invited into a world of enchanted 
mumbo-jumbo. But so far from there being any reference to what we 
might understand by religious instruction, Lucius simply receives a list 
of (expensive) things he has to buy, presumably the robes he later wears 
and the food for the banquets; and later receives whispered instructions 
{mandata) from Mithras (the priest), which cannot be divulged, but, to 
judge from the word itself, are likely to have been specific ritual instruc- 
tions. 32 What can be divulged to the reader relates to abstinence from 
wine and certain foods. If we disregard the 'Mithras-liturgy' and the 
supposed Berlin Mithraic catechism as too doubtful to use as evidence, 
as we surely must (see §5.b below), there is no other evidence for instruc- 
tion in Mithraism. As for the cult of Cybele, Graillot (1912, 537) cites 



29 E.g. 18.1: ek T0|j,7idvoi) fSePpcoKd...; 22. 1: Gapperte nvaxai It must be admitted 

that Firmicus Maternus' information about the cult of Cybele and Attis was evidently 
very sketchy. 

30 Apart from Winkler 1985, see e.g. J.L. Penwill, Ambages reciprocae: Reviewing Apu- 
leius' Metamorphoses, Ramus 19 (1990) 1—25. 

31 Litteris ignorabilibus praenotatos, partim figuris cuiusce modi animalium concepti sermonis 
eompendiosa verba suggerentes, partim nodosis et in modum rotae tortuosis capreolatimque condensis 
apicibus a curiositate prqfanarum lectione munita, "(books) inscribed with unknown characters. 
Some used the shapes of all sorts of animals to represent abridged expressions of 
liturgical language; in others, ends of the letters were knotted and curved like wheels 
or interwoven like vine-tendrils to protect their meaning from the curiosity of the 
uninitiated" (tr. JA. Hanson). 

52 Cf. Plautus, Capt. 343; Livy 22.22.16; 40.24.4; Tacitus, Hist. 4.49.2 (all specific 
instructions). The best parallels I can find to the use of the word in a religious con- 
text (our passage is not listed in TLL; in the apparently promising passage Ovid, Pont. 
2.2.121—4 Richmond mandata means 'request') are Ovid, Fasti 4.193f mandati memores 
(Cybele to the Muses); Met. 8.809f (an anonymous Oread communicates Ceres' instruc- 
tions to Fames): refert mandata deae, cf. 82l:functaque mandate; 14.829-31 (Hera to Hersilia 
via Iris); and some Christian texts, e.g. Tertullian, De patient. 15.2 Borleffs; Augustine, 
De civ. Dei 20.26, p. 749.12 Dombart-Kalb. 



216 CHAPTER FOUR 

a passage from Augustine which he claimed might be evidence of some 
sort of instruction: 

At enim ilia omnia quae antiquitus de vita deorum moribusque conscripta sunt, longe 
aliter sunt intelligenda atque interpretanda sapientibus. Ita vero in templis populis 
congregatis recitari huiuscemodi salubres interpretationes heri et nudiustertius audi- 
vimusP 

Augustine, Epist. 91.5 (Migne, PL 33, col. 315) 

However the continuation of the passage makes it clear that the refer- 
ence is to the cult of Capitoline Jupiter and other 'mainstream' cults 
since it refers to all manner of recitations of myths, including epic, 
plays and other public performances. 

The primary means whereby the lore of the mystery-component 
of the oriental cults was communicated seems to have been through 
direct participation in ritual. 34 That is certainly the impression given 
by Apuleius, Met. 11.19, 23-4 and 28. For what his evidence is worth, 
Firmicus Maternus likewise implies that cr6|i|3oA,cx were communicated 
for the first time in the course of rituals. 35 In the wider contexts of 
mystery- initiations, it is well-known that the words opyiov and opyia 
could be used concretely to mean 'sacred objects viewed at a moment 
of revelation' and thus as a synonym of ot)|i(3oA,ov or cr6u|3oA,oc used 
of physical tokens rather than utterances. 36 The Eleusinian mysteries 
showed the initiate "simple things that at the same time had a profound 
significance". 37 Such pointers strongly suggest that, by contrast with the 
Christians, that there was little or no attempt to fix dogma or teaching 
as part of a struggle for control within the cult. It was this absence of 



33 "All the old stories about the gods and their way of life have long since been 
reinterpreted, that is allegorised, by those with learning. In my young days, and even 
before, one could hear such revised interpretations being recited at temples when the 
people were assembled there." 

34 This process has been theorised by Bell 1992, 69—142; also the wider notion of 
'performativity' developed by the gender theorist Judith Butler as "a regularized and 
constrained repetition of norms [which] enables the subject": Bodies that Matter (New 
York 1993) 95. 

fa Habent enim propria signa, propria responsa, quae illis in istorum sacrilegiorum coetibus diaboli 
tradidit disciplina, "(The pagans) have special pass-words and appropriate responses which 
the Devil's instruction has communicated to them in their sacrilegious assemblies": 
De errore 18.1. 

36 A. Henrichs, Die Maenaden von Milet, Z PE 4 (1969) 223-41 at 225-9; Parker 
and IxauaTO;iot>Xot) 2004, lOf. 

!/ Clinton 2007, 355. The 'simple things' included of course the return of a lost 
daughter to her mother, a goddess in suffering, joy that accompanies the appearance 
of grain, the grain that is Ploutos . . . 



THE RITUAL SYSTEMS 217 

normative belief that ensured that there would be nothing resembling 
heresy; rather we should take it that there was a wide range of local 
and even individual variation. While we may allow that this afforded 
an agreeable flexibility that we should perhaps emulate in our studies 
of these matters, it does of course increase the danger of circularity 
when we attempt to specify the content of any given oriental cult. 38 

For all these reasons, we may take it that there was in fact a great 
variety of rituals in these religious movements, of which our sources give 
us hardly an inkling. Their basic dramaturgy owed something, but by no 
means everything, to the wider generative grammar of pagan festival; 
yet in many ways it was precisely their rituals that made them distinctive 
within the context of civic religious practice. The paradigmatic ritual 
of these cults was of course initiation, to which I now turn. 

a. Initiation 

tooto 8f| eOe^ov 8r|A,o , uv to tcov iiixyrnpicov xcov8e 
e7uraY|ia, to |if| eiapepeiv elq |if| |ie|ii)r||jivoi)<;, cb<; otjk 
eKcpepov eKeivo 6v, aneine 8r|?un)v 7tpo<; aXkov to 6eiov, 
otcp |if| Koci ai)%$> i8eiv evro%r\xm. 3 ' 3 

Plotinus, Enneads 6.9.11, p. 187.1—4 Brehier 

koci 7iaaa xtktir\ 7tpo<; xov k6<j|iov r\\\uq kou 7tpo<; xovq 
Qeovq fy)vanxeiv eQeXei. i0 

Sallustius, 7tepi 0ecov Kai koc|io\) 4.6, p. 7 
Rochefort = p. 6 1. 25f. Nock 



38 The point nowadays hardly needs making, but we might cite, for example, Isidore's 
four Hymns to Isis from Medinet Madi in the Fayyum of early I a , which are close 
to, but far from identical with, the tradition apparently emanating from Memphis (SB 
8138-41 = SEG 8: 548-51. The best text, with French transl, is E. Bernand, Inscrip- 
tions metriques de I'Egypte grko-romaine. Annales litt. de l'Universite de Besancon 98 [Paris 
1969] 631—52 no. 175 = Totti nos. 21—24; cf. also the commentary, unfortunately 
in ignorance of Bernand's text, by Vera Vanderlip (Vanderlip 1972). We might also 
think of the Prayer of Cascelia, found together with a number of other extraneous 
(i.e. non-Mithraic) monuments in the mithraeum of the Castra Peregrinorum beneath 
S. Stefano Rotondo on the Caelian, though I personally see no reason to associate it 
with any identifiable cult: AE 1980: 51; G. Mussies, Cascelia's Prayer, in Bianchi and 
Vermaseren 1982, 156-67. 

39 "'Pnij J s me intention of the command given in the mysteries here not to divulge 
(anything) to the uninitiated; (such a rule is proper, since) the divine can only divulge 
itself if one undertakes never to reveal it to someone who has not had the good fortune 
to behold it for himself." 

4(1 "Every initiation tends also to bring us closer to the kosmos and the gods." 



218 CHAPTER FOUR 

[Mystics] act as providers of evidence as to the truth of the 
supernatural for the body of believers that are not 
endowed with mystical capacity. 

R. Firth, Essays on Social Organization and Values 

(London 1964) 300 

It is difficult to deal briefly with a complex topic so charged in antiquity 
with different symbolic meanings and connotations, and about which 
so much has been written in the modern period. 41 My account here is 
based squarely on my reading of Apuleius, Met. 11.19-30. As I have 
explained several times, I take these pages, although they ostensibly 
relate only to the cult of Isis, as an essentially trustworthy and in 
principle generalisable account of mystery initiation in the mid-second 
century AD. 42 From this perspective, initiation is to be seen as a rite 
of transition whose main feature was a fictive death and rebirth to a 
new life. 43 In so 'dying', the initiand is able to quit his real temporal 
existence and gain access to the mythical time of divinity. He is thus 
able to participate in the cosmogony and be united with the divine 
creator, the demiurge responsible for the order displayed by the cosmos. 
Rebirth involves becoming a new being, a metamorphosis that makes 
it possible for the initiand to enter into a new reality (Fick 1992), set in 



41 See the bibliography assembled by A. Moreau in idem 1992b, 297—305. 

12 I thus accept the main lines of the commentary of J. Gwyn Griffiths (Griffiths 
1975) and fully endorse the reading of Fick 1992. 

" Eliade 1959, 12: "On comprend generalement par initiation un ensemble de rites 
et d'enseignements oraux qui poursuivent la modification radicale du status religieux et 
social du sujet a initier. Philosophiquement parlant, l'initiation equivaut a une mutation 
ontologique du regime existentiel" (the formulation is repeated in idem, L'initiation et 
le monde moderne, in Bleeker 1965, 1); cf. CJ. Bleeker, Some Introductory Remarks 
on the Significance of Initiation, in idem 1965, 14-20. Ries 1986c, 27: "Par l'initiation 
s'opere un passage, d'un etat a un autre, d'un stade de vie a un stade nouveau, d'un 
genre de vie a un autre genre de vie. Ce passage se fait en vue de la realisation 
d'un type de comportements qui seront des donnees essentielles dans la vie et dans 
l'existence de l'initie . . . Deux elements semblent . . . essentiels: d'abord l'introduction 
dans une communauti, ensuite l'introduction dans une monde de vakurs spirituelles en vie 
d'une vie ou d'une mission" . A good up-dating of A. Van Gennep's ideas can be found 
in N. Belmont, La notion du rite de passage, in Centilivres and Hinard 1986, 1—15. 
Lhere are also a number of good articles on the topic in Bianchi 1986b, particularly: 
J.Y. Pentikainen, Lransition-rites, pp. 1—24; U. Bianchi, Some Observations on the 
Typology of 'Passage', pp. 45—61; N. Gasbarro, La grammatica dei riti di passagio, 
pp. 205-23. With P. Bourdieu, Les rites comme actes d'institution, in Centilivres and 
Hinard 1986, 206ff = idem 2001, 175—86 (though he makes no mention of what Eliade 
had already said fifty years earlier), I would really prefer the term 'rites of institution'. 
Despite its title, Y. Dacosta, Initiations et societes secretes dans I'Antiquite greco-romaine (Paris 
1991) is disappointing. 



THE RITUAL SYSTEMS 219 

a new physical or imaginary realm, where he enters into communion 
with the divinity to whom he is bound, a communion that attests to 
the depth of his commitment. The ritual also imparts secret teaching 
about the 'true' order of the cosmos. 44 Thanks to this knowledge, whose 
value was increased by rule of non-divulgence, 45 the initiand is able 
to share in overcoming chaos, represented by the fictitious death. 46 All 
this seems to have involved a metaphorical journey from ignorance to 
'full' knowledge, consisting in an ecstatic union with or vision of the 
deity. This journey could also be represented as tripartite, consisting 
of toe Spcoiievoc, the performance of the prescribed ritual, Sencvoixevcx, 
viewing the sacred objects or symbols, and Xeyoueva, uttering the sacred 
responses, the cr6|a,poA,cx (Henrichs 1998). 

Initiation gave the adherents of these cults to understand that, though 
humans are subject to Destiny, they can evade it through the aid of 
a salvific deity. This was the main secret that might not be divulged. 
The socio-political elite had nothing against all this. The restriction of 
such supposed knowledge to a smallish group was to their advantage, 
since most people continued to believe in the force of Destiny, i.e. the 
way things are. 47 Moreover escaping from Destiny, especially on the 
conditions offered, was no major threat to public order. It was all a 
matter of understanding how the implicitly centrifugal tendencies of 
small private cults were to be recuperated. One example might be 
the way in which the taurobolium/criobolium, though ostensibly a foreign 
ritual, served as an expression of loyalty to the domus divina; another the 



44 On secrecy as a resort, cf. Giebel 1990; 13; Turcan 1992a, 225; S. Montiglio, 
Siknce in the Land of Logos (Princeton 2000) 25; F. Graf, Confession, Secrecy and Ancient 
Societies, in Luchesi and von Stuckrad 2004, 259-71; Clinton 2007, 342-4. 

45 Pausanias relates that a man who had not been admitted to the ranks of 
Kaxa[3aivovxe(; at the temple of Isis at Tithorea in Phthiotis nevertheless entered the 
aSrjxov, saw the dead face to face, went back home, revealed what he had seen, and fell 
dead (10.32.17). Pausanias had heard a similar story about the temple of Isis at Coptos 
in Upper Egypt: in this case the man had allegedly been bribed by the Roman governor 
to penetrate the mystery, and he too is reported ocotikoi TeXemfjciai (ibid. 32.17) — a 
typical item of 'resistance' lore. On the early imperial 'voice oracle' at Coptos, see e.g. 
Apuleius, Met. 1 1.28: per adyta Coptica, with Frankfurter 1998, 150f, 156 etc. 

46 Eliade 1959, 17; cf. Brelich 1965 (an article that is fundamental to my view here); 
also R. Guenon, Aperfus sur I'initiation (Paris 1985). On Eliade 's curious dependence on 
Frazer's notion of comparativism, note N. Spineto, Le comparatisme de Mircea Eliade, 
in Boespflug and Dunand 1997, 93-108. 

47 Cf. C. Weiss, Deaefata nascentibus canunt, in H. Froning, T. Holscher and H. Mielsch 
(eds.), Kotinos: Festschrift fur Erika Simon (Mainz 1992) 366—74. 



220 CHAPTER FOUR 

patterning of certain Mithraic rituals on the model of the patron-client 
relationship. 48 

Initiation inaugurates a new life. The transition from one to another 
permits one to leave behind the negative aspects of one's previous life, 
and to make one's appearance as a new being beyond biological death. 
As far as the mysteries were concerned, as we saw in the fragment of 
Plutarch cited in Chap. 2.2 above, where xeXevxav is equated with 
xzkziaQai, 'true' death occurred so to speak at the moment of initiation 
(cf. Alvar 1993a, 132 and 139). This new being can be born thanks to 
a vision of another world that is the specific construction of the cult 
in question. There is no reason why this world should be opposed to 
the dominant values of its day, though of course it emphasises those 
values that are considered essential in pursuing the objectives set out, 
implicitly or explicitly, at initiation (Eliade 1959, 12). Nor need these 
values be identified with the sacral, since there is no reason to suppose 
that the promises of the mysteries were confined to well-being conceived 
in a purely religious framework. 49 As I have already stressed, the notion 
of ocoxtipia was anchored in mundane, physical well-being. Indeed, as 
its symbolism indicates, initiation was often linked with the desire for 
children (Meslin 1985). 

An important aspect of initiation is that it takes place away from 
everyday life. 50 To be accepted as true or effective, the metamorphosis 
has to be experienced in a strange, a different, place. 51 It may be con- 
sidered to be located somewhere between the human and the divine. 
Access to it must be restricted. Only one who has been instructed by 
the mystagogue can enter the 'womb' where the change took place; and 
the catechumen can in his turn become a father who leads the next 
generation to the maternal 'womb'. 52 Given the extraordinary nature 



18 Taurobolium: Rutter 1968, 233—38, though see R. Sierra, Taurobolio y el culto 
imperial en la Galia Narbonense, in Alvar, Blanquez and Wagner 1995, 201—14; 
Mithraism: Gordon 2001b; 2007, 402f. 

19 This has been amply demonstrated for the cult of the Mater Magna by Sfameni 
Gasparro 1985; here I differ from Ries 1986c, 27. 

30 I am here following up some ideas on the location of initiation first suggested to 
me by reading Meslin 1986 and J.G Simon, Initiation et espace sacre, in Ries 1986b, 
107-27. 

51 Cf. C. Calame, Prairies intouchees et jardins d'Aphrodite: espaces 'initiatiques' 
en Grece, in Moreau 1992b, 103-18. 

32 Meslin 1985 has argued for a close link between the location of initiation and 
the basic features of biological reproduction, suggested by the primeval or elemental 
imagery offered by Apuleius. Whether these associations were conscious to the minds 
of either mystagogues or initiands is another question. The imagery of fatherhood 



THE RITUAL SYSTEMS 221 

of what takes places on this spot, it must be aesthetically effective so 
as to be able to give an impression of primordial chaos and terrifying 
darkness, 'cosmic night'. Primordial or mythical time also makes an 
appearance as part of the demiurgic process, marked off from real 
time by a sort of amnesia — symbolic death; and in this primordial time 
occur the journeys, 'to' the Underworld but also evidently 'to' heaven, 
we hear of in Apuleius: 

accessi confinium mortis et, calcato Proserpinae limine, per omnia vectus elementa 
remeavi; nocte media vidi solem candido coruscantem lumine; deos inferos et deos 
superos accessi coram et adoravi de proximo.™ 

Apuleius, Met. 11.23 

We might also think of the solar 'journeys' that Beck has argued were 
practised in Mithraism, mini-processions that took place in the artificial 
cave of the temple. 54 At the same time, the initiate must also return to 
this same locus as a new being, so that the staging had to be sufficiently 
adaptable to represent the light of the new life he or she has managed 
finally to attain. 55 

b. Sacrifice 

Blut ist ein besondrer Saft. 

Goethe, Faust I, 1. 1740 

Wohin ist Gott, rief er, ich will es euch sagen! 
Wir haben ihn getodtet — ihr und ich! 
Wir Alle sind seine Morder . . . Gott ist todt! Gott 
bleibt todt! Und wir haben ihn getodtet! 

Friedrich Nietzsche, Diefrbhliche Wissenschqft 
Buch 3, no. 125 (eds. Colli and Montinari, Berlin 

1973, p. 158f.) 



seems to have been extensively employed only in the cult of Mithras, and to a much 
lesser extent in the cult of Isis. Mater is sometimes found as an honorific title for women 
who provided funds for collegia; but that is a completely different matter. 

53 "I came to the boundary of death and, having trodden the threshold of Proserpina, 
I travelled through all the elements and returned. In the middle of the night I saw the 
sun flashing with bright light. I came face to face with the gods below and the gods 
above and paid reverence to them from close at hand" (tr. J.A. Hanson). 

54 Beck 2000, 154-67 = 2004a, 64-77, but see §5. a below for some doubts. 

55 The language of the passage, like that of Plutarch frg. 178 Sandbach, perhaps 
suggests that much, perhaps all, was in the mind, so that very little in the way of stag- 
ing may have been required, except for lighting, cf A. Motte, Nuit et lumiere dans les 
mysteres d'Eleusis, in J. Ries and CM. Ternes (eds.), Symbolisme et experience de la lumiere 
dans les grandes religions: Actes du collogue de Luxembourg, 29—21 mars 1996. Homo Religiosus, 
series 2, 1 (Turnhout 2002) 91-104. 



222 CHAPTER FOUR 

vupie, toe xs^TI M-od dvoi^en;, Kai to at6|ia |iot> 
dvayyeXei tpv aiveaiv gov oti ei r\QeXr\aaq 
Qvaiav, e8a>Ka dv- 6XoKaut(0|iata otjk 
ex)8oicriaei(;. Svaia tco 6eco 7ive{)|ia 
awtetpi|i|ievov, Kap8iav cwtetpi|i|ievTiv Kai 
teta7ieivco|ievr|v 6 Qebq cuk e^o\)8ev(oaei. 

Psalm 51 w. 17—19 (Septuagint version) 

One obtains divine favour by trying to please god, and there are few 
more expressive ways of doing that than offering something one needs 
oneself: food. 06 Even if the honorand neglects to bestow a return gift, 
the gods must be pleasured, since neither they nor their worshippers 
could see as clearly as Socrates the degree to which piety (baw%y\q) is 
about making deals (eu7topiicf| xzxvr]). 3 ' This is not the place for a gen- 
eral discussion of sacrifice, and I want simply to resume a few points 
of greatest importance for the oriental cults. 58 Sacrifice in these cults 
corresponded, albeit on a much reduced scale, to the wider practice we 
call sacrificial euergetism. 59 So far as we know, their sacrificial praxis 



56 Porphyry, De abstin. 2.24 offers three reasons for sacrifice: to honour the gods, 
to return them thanks and to get from them some desired benefit (he omits divina- 
tory sacrifice). Saloustios/Sallustius, ilepi 0ecov Kai koc^ou 16.1 notes that it is right 
and proper for us to offer the first fruits of what has been given to us to those who 
bestowed it in the first place: of our wealth we do this through making votives (8i' 
dva0a)j,dx(ov), of our bodies "by means of a lock of hair" (5ia kout|<;), of [our?] life 
"through sacrifices" (8ia 0t)aiwv). For a modern analysis, cf. C. Riviere, Approches 
comparatives du sacrifice, in Boespflug and Dunand 1997, 279-89. 

31 Plato, Euthyphro 14b— e, cited by Burkert 1996, 135. Nearly 800 years later Salous- 
tios/Sallustius affirms that, since to GeTov in itself needs nothing from us, the honours 
we pay divinity are xx\c, fmetepcxQ wiptXeiaq evekck (Iispi 0eSv Kai Koauot) 15.1). Of 
course this line is common enough among philosophers, who were generally more in 
favour of ethical improvement than hecatombs: e.g. Cicero, De nat. dear. 2.7 If with Pease 
ad loc; Philostratus, Vit. Apollon. 1.1 Of. The neo-Platonist and Egyptian priest Iamblichus 
elevates the tone still further by saying that we ought not to permit sacrifices either in 
honour of the gods, nor to give thanks, nor as votives, nor as first-fruits, since all these 
are vulgar misapprehensions. Nothing counts but the gods' absolute superiority and 
their role as transcendent causes of things (De myster. aegypt. 5.5-7). In this connection 
it is worth repeating that public sacrifices seem already to have ceased in most cities 
by the time Julian became emperor: Bradbury 1995; Stroumsa 2005, 105—43. 

58 In his introduction to Neusch c. 1994, the editor stresses that there is far too much 
variety of practice to permit universalising definitions. The contributing authors express 
these differences each according to his epistemological and confessional inclination. 
As for social-anthropological views, we may highlight that of VW. Turner, Sacrifice 
as Quintessential Process. Prophylaxis or Abandonment?, History of Religion 16 (1977) 
189-215. J.H.M. Beattie, On understanding Sacrifice, in Bourdillon (and Fortes) 1979, 
1—25 misguidedly excludes a priori all non-animal sacrifice. 

59 Gf. S. Pierce, Death, Revelry and thysia, ClAnt 12 (1993) 219-66; Schmitt Pantel 
1999; P Veyne, Bread and Circuses: Historical Sociology and Political Pluralism, tr. D. Sul- 



THE RITUAL SYSTEMS 223 

owed nothing to their earlier developmental phases, though claims 
have been made to the contrary, particularly in relation to the tauro-/ 
criobolium. The most important single fact is that, in common with 
Graeco-Roman civic sacrifice, praxis centred upon the altar as the 
point of communication with the other world. 60 

Civic sacrifice was understood as part of a complex of actions, with a 
grammar of its own, whose purpose was to render divinity propitious. 61 
As Jean Rudhardt has stressed in the case of Greek sacrifice, its elements 
cannot be rigidly classified. 62 Nevertheless we can distinguish between 
different modalities, in the case of liquids, say, between libation and 
aspersion, when the latter is intended as an offering and not simply as 
a means of purification. 63 I certainly find it tempting to think in terms 
of the relationships set up by different actions: aspersion, for example, 
has a collective connotation, inasmuch as it links all those on whom 
the water falls, whereas libation simply establishes a relationship with 
the recipient, on the implicit assumption that it is up to the divinity to 
redistribute its beneficial effects. 



livan (New York and London 1990, orig. ed. 1976). T.D. McCreight, Sacrifical Ritual 
in Apuleius' Metamorphoses, in H. Hofmann (ed.), Groningen Colloquia on the Novel 5 
(Groningen 1993) 31—61 makes an interesting attempt to integrate Apuleius' accounts 
of sacrifice into the wider Roman system. The role of the mysteries in the evolution 
of representations of sacrifice is over-simplified by Eisner 1995, 190—245: tradition- 
ally the gods were the recipients of sacrifice; Mithras performs one himself; whereas 
Christ is himself the victim. This account however ignores both the traditional Greek 
iconography of gods offering libation, the subject of a well-known debate in Germany, 
and the equally traditional role of Nike/Victoria as bull-sacrificer, which is directly 
relevant to the case of Mithras. 

60 J.-L. Durand, Betes grecques, in Detienne and Vernant 1979, 133—81 at 139. 

61 Cf. H. Foley, Ritual Irony: Poetry and Sacrifice in Euripides (Ithaca and London 1986) 
24-39; J.-P. Vernant, A General Theory of Sacrifice and the Slaying of the Victim 
in the Greek thysia, in idem, Mortals and Immortals: Collected Essays (F.I. Zeitlin, ed.) 
(Princeton 1991) 290-302. 

1)2 Rudhardt 1958, 213. The terms 0rjoia and cntov5r| have been analysed by 
J. Casaubon, Recherches sur le vocabulaire des sacrifices en grec des origines a la fin de I'epoque 
classique (Aix-en-Provence 1966) 126—39 and 256. In his view the first denotes any con- 
secrated offering, animal or vegetable, some of which is destroyed and the remainder 
consumed; the second is a general term applicable to any libation, irrespective of the 
nature of the liquid involved, the circumstances and the recipient. 

63 Cf. Stengel 1910, 34-39; 1920, 103-05; Eitrem 1915, 76-83. J. Maisonneuve, 
Les rituels. Que sais-je? 2425 (Paris 1988) 30 distinguishes between ablution, confession 
and penitence, inversion and expulsion as sub-categories of purificatory rites, whose 
purpose is to avoid the negative consequences of contact with impurity. 



224 CHAPTER FOUR 

The Greeks and Romans saw no qualitative difference between animal 
and vegetable offerings, even though in principle they might have been 
understood as two separate expressive systems (of course this did not 
apply to other forms of differentiation such as methods of cooking). 64 
As L. Ziehen put it: 

Sicher aber ist, daG die unblutigen Opfer wahrend des ganzen Altertums 
eine hervorragende Stellung behaupteten, nicht nur weil sie zu den 
altesten Opfern gehorten und durch diese Tradition geschiitzt waren, 
sondern vor allem deshalb, weil die vegetarische Nahrung selbst bei 
den Griechen und Romern, wie iibrigens schlieGlich auch heute noch, 
vorherrschend blieb. 65 

As for animal-sacrifice, while votives might well be private affairs, in 
general sacrifice and libation had a social dimension. Even 'private' 
sacrifices were generally undertaken within the family context. More 
leeway was available in the case of oracular consultations, for example 
at consultations of the oracle of Trophonius in Boeotia, conspiracy, or 



''"' The Paris School has always tended to neglect the existence of vegetable offerings, 
e.g. J.-L. Durand, Sacrificare, dividere, ripartire, in Grottanelli and Parise 1988, 193: 

"II sacrificio consiste nell' uccidere un animale, secondo certe procedure adeguate, 

per mangiarlo". Yet, as everyone knows, Theophrastus believed that they were the 
original type of offering, just as humans were originally vegetarian (Porphyry, De abstin. 
2.5.2), cf. G. Bodei Giglioni, Come gli uomini devennero malvagi: sviluppo della civilta, 
alimentazione e sacrificio in Teofrasto, RSI 103 (1991) 5—32 (though her particular 
argument, that these ideas are not Theophrastan, is quite unconvincing); Bouffartigue 
and Patillon 1979, 67—71. The importance of bloodless offerings in Greek offering 
praxis was explored already by C . A. Lobeck, De Graecorum placentis sacris (Konigsberg 
1818); and 1829, 1050-85 (the famous 'Pemmatologia sacra'), 1082-85; and they 
received considerable attention in the work of the 'philological' school: O. Band, 
Das Attische Demeter-Kore-Fest der Epikleidia, Programm der Margarethenschule Berlin 
(Berlin 1887) 4fF.; G. Hock, Grieehische Weihegebrduche (Diss. Munchen 1904) (Wiirzburg 
1905); Stengel 1910, 7£, 66-72, 130-32; Wyss 1914; Eitrem 1915, 261-80 (I exclude 
the discussion of salt); C. Mayer, Das 01 im Kultus der Griechen (Diss. Wiirzburg 1917); 
Stengel 1920, 98-103. 

65 L. Ziehen, s.v. Opfer, RE 18 (1939) 579-627 at 582; the idea is repeated from 

0. Kern, Die Religion der Griechen (Berlin 1926—38) 1: 156f, who cites the female figure 
offering fruits and wine as part of the bull-sacrifice on one of the long sides of the 
Poros-stone Late-Minoan III sarcophagus from the chamber-tomb near the palace of 
Haghia Triada (c. 1400 BC); cf. S. Marinatos, Crete and Mycenae (London 1960) 15 If 
with the colour photo pi. XXVIII (facing p. 66); M.P Nilsson, The Minoan-Mycenaean 
Religion and its Survival in Greek Religion} Skrifter utgivna av kungl. humanistiska vetens- 
kapsamfundet i Lund 9 (Lund 1950) 426-443 at 432f; a b/w photo in Nilsson, GGR 

1, pi. 10.1. Cf. Kirk 1981, 77: "One of the remarkable facts about Greek divine cult 
in the late Bronze Age is that . . . burnt animal sacrifices were comparatively rare, and 
the regular means of worship was through libation and the presentation of vegetable 
substances (cereals, honey and the like), on stone 'tables of offering'". 



THE RITUAL SYSTEMS 225 

magical practice, though in no case should we lose sight of the link 
with private, and in some cases even public practice. 66 

It used to be an article of faith that official or public Graeco-Roman 
religion afforded no opportunity for profound religious emotion com- 
parable to that of the mystery-component of the oriental cults (or for 
that matter Christianity — usually the implied comparandum). It is now 
routinely accepted that there was a place in all of them both for the 
magical-comminatory and the submissive-religious mode. 6 ' But my main 
interest here is to emphasise that religious feelings can be manifested 
in all kinds of ways, not necessarily hidden away inside the individual; 
and that any public ritual, perhaps even especially rituals organised by 



l>l> Cf. Malaise 1986a, 92. Trophonius: P. Bonnechere, Trophonius de Lebadee: cultes et 
mythe d'une cite beotienne. RGRW 150 (Leyden 2000); oath: Lollianos, Phoinikika Bl recto 
et verso: A. Henrichs, Lollianos, Phoinikika: Fragmenta eines neuen Romans, £PE 4 
(1969) 205-15 at 206f. = idem, Die Phoinikika des Lollianos. Papyrologische Texte und 
Abhandlungen 14 (Bonn 1972); idem, Human Sacrifice in Greek religion, in Rudhardt 
and Reverdin 1981, 195-235 at 228f., but the interpretation of G.N. Sandy, Notes on 
Lollianus' Phoenicica, AJPh 100.3 (1979) 367-76 is to be preferred to Henrichs' 'Mys- 
terienroman' theory; magic: of many texts, Horace, Epode 5 with the commentary by 
L.C. Watson, A Commentary on Horace's Epodes (Oxford 2003) 174—250. 

67 Some people are still inclined to believe that depth of religiosity is correlated 
with social complexity. Let me here reiterate my conviction that a variety of religious 
expressions can co-exist in complementary fashion, whether or not those higher up in 
the social order persist in thinking that they are incompatible. I do not here wish to 
enter into the discussion on the relative religiosity of those who live in states and in 
acephalous or non-state societies. Elsewhere in this book I have argued that religion 
is best thought of as closely connected to the existence of the state; here I want to 
stress that elevated forms of religiosity can also be found in acephalous societies (cf. 
p. 208). I find it odd that it is usually those who claim that religion is consubstantial with 
humanity who also believe that 'primitives' have no access to ethical-spiritual experience, 
as though such experience passes in a single leap from a few primitive forms to truly 
complex spirituality. All this is simply the precipitate of nineteenth-century evolutionary 
thinking. It is still to be found in the work of someone like A. Cahn, Le sacrifice dans 
l'hindouisme, in Neusch c. 1994, 201, who understands the shift from comminatory 
to submissive sacrifice as one from a magico-ritual level to an ethical-spiritual one, as 
though it were obvious that the one must necessarily be later than the other, and, of 
course, higher in the grand scheme of things. I believe in their possible co-existence, 
since ritualisation is no less a feature of the supposed second (ethical-spiritual) level 
than of the first, and both ethics and spirituality can attain the same quality in the 
supposed 'magico-religious' phase as in the 'ethico-spirituaP one. Given my rejection 
of evolutionism, it will be obvious why I prefer the notion of mode, as in the main 
text. It has the additional advantage of being easier to accept for those who do not 
hold that most divine-human relations are a result of institutionalised religion. On all 
this, I do urge the reader, even if he or she wants nothing to do with biologism, to 
look at Burkert 1996; also idem, Ritual between Ethology and Post-Modern Aspects: 
Philological-historical Notes, in Stavrianopoulou 2006, 23—36. 



226 CHAPTER FOUR 

the collectivity, can arouse deep religious feelings even though that is 
not their main intention. 

With Marcel Mauss, I would thus argue that sacrifice involves the 
idea that material and (supposed) spiritual energy can be transformed 
through ritual into a concrete, apprehensible physical benefit. In order 
to avoid any direct contact between sacrificant and deity, such trans- 
formation requires a mediator, who must afterwards disappear. 68 At 
any rate, whatever the precise nature of sacrifice, its functions can be 
analytically distinguished (though not, generally speaking, by the actors) 
as apotropaic, purificatory, propitiatory and regenerative. In the course 
of human history only a few societies have developed the ideological 
conditions necessary to construct a regular theory of sacrifice. 69 One 
example would be Theophrastus' argument in llepl zvaz^ziac, that true 
sacrifice consists solely of vegetable offerings, which derive from our 
common mother, the earth: only then, and only on condition that they 
are offered in the right frame of mind, shall we be considered worthy, at 
the end of our lives, to behold to crouTtcxv yevoc; xcov ev ovpocvcp 9ecov. 70 
The reward for correct sacrificial practice is thus projected into the next 
world. Only once this step has been taken, can the direct link between 
offering and return be broken. And only then can the focus be upon 
the qualitative value of the offering rather than the quantitative. 

In Chapter 2, I suggested that the ancient world had arrived a pre- 
cisely such a stage at the time the oriental cults reached their apogee 
during the second century AD. That was my reason for arguing in 
Chapter 3 that we can properly speak of their systems of values. The 
alteration in the significance of sacrifice is a symptom of a paradigm- 
shift with regard to the conceptualisation of the relation between the 
divine world and this. I see this shift as dialectically related to the cults' 



68 H. Hubert and M. Mauss, Sacrifice: its Mature and Function. Tr. W.D. Hall (London 
1964, orig. ed. 1899); cf. M. Mauss, Essai sur le sacrifice, in Oeuvres I (C. Levi-Strauss, 
ed.) (Paris 1968). Mauss, as a good Durkheimian, argued that sacrifice is a medium 
whereby the profane communicates with the sacred via a victim, i.e. something con- 
secrated, that is destroyed in the course of the ceremony, so as to avoid any direct 
contact with the sacred. 

69 I do not count the Hesiodic myth of Pandora as a theory in the required sense, 
cf. Csapo 2005, 251—62. The most elaborate traditional theory of sacrifice must be 
the Mimamsa Brahmin theory of apurva. 

70 Porphyry, De abstm. 2.32. If, cf. Bouffartigue and Patillon 1979, 17-29, 51-58; 
D. Obbink, The Origin of Greek Sacrifice. Theophrastus on Religion and Cultural 
History, in W. Fortenbaugh and R.W. Sharpies (eds.), Theophrastean Studies. On Natural 
Science, Physics, and Metaphysics, Ethics, Religion and Rhetoric. Rutgers University Studies in 
Classical Humanities 3 (New Brunswick 1988) 272-95. 



THE RITUAL SYSTEMS 227 

morality: it made the new moral perceptions more attractive, and was 
in turn fostered by them.' 1 

I take the absence of documentation that could decisively resolve such 
issues as an indication that the religious field was indeed in the course 
of a transition of the sort I have tried to outline in this discussion of 
sacrifice. 72 However, as I have said before, I do also acknowledge that 
the sacrificial regime continued to maintain the existence of a 'magical' 
relationship between the community and the victim, which legitimated 
the consumption of the latter. From that point of view red meat and 
fat surely remained important. But the most charged element in this 
symbolic universe was surely not the meat but blood, that special 'juice' 
that was recognised by sacrificants as being able to circulate the desires 
and wishes of mankind into the other world and out of it again. And, 
in our special case, blood bundled purificatory power, regeneration, 
physical and moral strength, everything in fact that one required for 
initiation. For in that rite, more than in any other, life and death, mate- 
rial existence and spiritual ardour, cosmic regeneration and personal 
immortality were fused inextricably together. 73 

c. Commensality 

ox) yap oaep gov 
eiXajtivai Svptoiaiv iv' cu 7ip(0Tr| 7ti)|a6Vt:r| xz 
'Ecmri &px6|ievo<; cntev8ei |a.e^ir|8ea oivov. 

Hymn. Horn. 29 to Hestia, 11. 4-6 Allen 

Man amiisierte sich gottlich in [den] Tempeln, bei 
[den] Festspielen, Mysterien; da schmiickte man das 
Haus mit Blumen, da gab es feieriich holde Tanze, 
da lagerte man sich zu freudigen Mahlen . . . wo nicht 
gar zu noch siiBeren Geniissen. 

H. Heine, Elementargeister (1835—7) — Samtliche 
Schriften (ed. K. Briegleb) (Munich 1996 3 ) 3: 685. 



" Cf. now the complementary arguments of Stroumsa 2005, 105—44. 

' 2 Cf. D. Frankfurter, On Sacrifice and its Residues: Processing the Potent Body, in 
Luchesi and von Stuckrad 2004, 51 1—53, arguing for a view of sacrifice that acknowl- 
edges the reality of what passes in ritual from the other world into this one. 

''' On blood, see still the extraordinary chapter by Eitrem 1915, 416—60; the second 
paragraph begins wonderfully: "Es gibt noch Volker, die frisches Menschenblut als 
Nahrungsmittel . . . genieBen"; on blood in the cult of the Mater Magna, cf. Sfameni 
Gasparro 1983. Kirk 1981, 52f. is rightly sceptical of the idea of the identity of sac- 
rificant and victim, which sustains so much modern theorizing about sacrifice. 



228 CHAPTER FOUR 

Once the libations have been poured and the victim sacrificed, the 
banquet is prepared by the cooks. A Christian like Clement of Alex- 
andria could claim that sacrifice was invented as an excuse to eat meat: 
oapK09ayicov 8' oium 7ipo(pdoei od 9i)aicxi xoiq 6cv9pamoiq £7UV£vor|VTca 
(Stromateis 7.6.32 — III, p. 24. 16f. Stahelin). Such a claim is absurd, as 
Clement himself half-admits, since it would imply that all meat eaten 
in antiquity was sacrificial, that only meat was sacrificed, and that sac- 
rifices invariably ended in a meal. None of these propositions is true: 
meat from hunting is an obvious exception to the first claim, as are 
holocaust or naval sacrifices to the third (though holocaust sacrifices are 
uncommon and limited to small animals such as piglets). Nevertheless 
a feast involving the consumption of meat often was the final act in 
the sacrificial process, as we can see from the numerous representa- 
tions on red-figure vases of Herakles roasting anXajx va on l° n g spits, 
in preparation for la grande bouffe. 1 * On the other hand, access to such 
feasts was usually restricted: although, as I have mentioned, the Archaic 
and Classical stoa was developed in order to provide shelter for large 
public banquets, and Hellenistic and Roman euergetes and foundations 
did sometimes finance similar occasions, 73 in the normal run of things 
only the priests and sacrificants, i.e. those who paid for the victim(s), 
and their families, were admitted to the god's table. 76 As a result, at 
least in Athens, there appears to have been a certain resentment of the 
lavishness of such private banquets. 77 

The sacralisation of the feast is hardly to be wondered at since 
food provides the energy required to live. However, the process can 
also been seen as one aspect of the wider regularisation of sacrifices 



'* Cf. Rudhardt 1958, 158; J. -E Vernant, A la table des hommes, in Detienne and Ver- 
nant 1979, 37-132 esp. 43-45; A. Motte, Le symbolisme des repas sacres en Grece, in 
Ries 1985, 157—71. Herakles cnrA-ayxvoTtTrn; (a theme entirely missing from F. Brommer's 
various collections): G. Rizza, Una nuova pelike a figure rosse e lo "splanchnoptes" di 
Styppax, Annuario della scuola Archaeologica di Atene 37-38 (1959-60) 321-45. 

73 G. Kuhn, Untersuchungen zur Funktion der Saulenhalle in archaischer und klas- 
sischer Zeit, JdAI 100 (1985) 169-317; Schmitt Pantel 1999; also earlier, e.g. Nikostra- 
tos son of Dieitrephes, several times general between 427—18 BC: cf. Aristophanes, 
Wasps 82 with D.M. MacDowell ad loc; idem, Nikostratos, CQn.s. 15 (1965) 41-51 
esp. 48-50. 

76 Mylonopoulos 2006, 79—83; K. Dunbabin, Ut graeco more biberetur. Greeks and 
Romans on the Dining-Couch, in Nielsen 1998, 81-101 at 82-89. 

77 As a comic topos, cf. Pherecrates, Automoloi frg. 23 K = 28 Kassel-Austin (vol. 7, 
p. 118); Menander, Dysc. 447-54 with E.W. Handley ad loc. (p. 214); also Aristotle, 
EM 1160al9: evusu §e tSv koivwvubv Si' fiSovnv SokoCcu yvyveoGcu, OiaacoxSv kcxi 
epaviCTmv (perhaps merely descriptive, but see G. Ramsauer ad loc, p. 546). 



THE RITUAL SYSTEMS 229 

and libations so as to form an annual ritual calendar that orchestrates, 
explicitly or implicitly, the more important festivals of the polls (whether 
Greek or Roman) and its sub-divisions. Being private institutions (if 
we exclude the Megalensia), the oriental cults had only an indirect rela- 
tion to the public sphere: their proper Sitz im Leben was the group of 
adherents. 78 Nevertheless they did have a public dimension (expressed 
for example in the Isidis navigium), a dimension that increased as they 
became socially more integrated into the ideological apparatus of the 
state, culminating in rituals and votives specifically intended to uphold 
the public welfare, particularly of the emperor. From that perspective, 
their rituals, sacrifices and banquets had implications at the level of 
the individual, the group and the state as a whole. 

We cannot decide for certain whether the ritual banquet was a recent 
innovation in the mysteries or in some sense 'original'. But there is 
some archaeological evidence in favour of the assumption that it was 
an early feature of these cults as they spread into the Mediterranean. 
Serapeum A at Delos for example, which was built c. 240 BC, has 
a dining-room of c. 40 m 2 next door to the temple, which would be 
enough to entertain a small number of people (either because there 
were no more in the group or because not everyone attended at the 
same time). Maiistas, the temple-aretalogist, describes the ensemble in 
his foundation-narrative: 

ceSev 0' oc|ia (3o\)^.o|ievoio 
pr|i8i(o<; Koci veux; ae^eto kocI 6i)6evT:e<; 
Pco|xoi koc! TC|i£vo<;, xexeXeaxo 8e 7idvta |ieXd0pcoi 
65 e8pavd te kA,i0|ioi te 0eoKA,f|"tcn)<; en\ Sauaq . . .' 9 

IG XI.4, 1299 = Totti no. 11 = RICIS 202/0101 11. 62-65 

The expression 9£okA,t|toi)c; k%\ SaTxaq surely indicates the sacral, 
perhaps even sacramental, character of these feasts. 80 Many invitations 
to such Serapis-banquets have been found on ancient rubbish-dumps 



/S Bleeker 1963, 233—35 draws up a scheme of four types of sacral meal, the last 
of which corresponds to my initiatory meal, which he treats as a sub-type of 'com- 
munion'. 

79 "Because you (i.e. Serapis) willed it so, the temple, and the incense-burning altars 
and the entire temenos were built with ease, and all the seats and couches were constructed 
in the big hall for the banquets to which the god summons (his followers)." 

80 Cf. Engelmann 1975, 42f.; MacMullen 1981, 37-48. Despite its merits, I think 
that Egelhaaf-Gaiser 2000, 300—29, in treating temple, collegium, synagogue, mystery 
and funerary meals together, fails to make the appropriate distinctions between different 
types of meals, and seems to believe they all had the same value. 



230 CHAPTER FOUR 

in Egypt, particularly at Oxyrhynchus; Tertullian, hyperbolic as ever, 
claims the cooking will be so lavish, and cause so much smoke, that 
the fire-brigade may have to be called {spartioli excitabuntur)} 1 We can 
certainly say that such meals were usual in all three of these cults: 
the common meal is one of the most effective means of integrating 
the members of a community and creating feelings of solidarity, thus 
providing a significant contrast to the world outside; it is also a central 
focus of funerary celebrations. 82 The mithraeum is indeed a perfect 
exemplification of this aim, with its podia, in the form of extended 
klinai, running down each side of the building, and the tauroctony the 
god performing his sacrifice, on the back wall. 

What of the symbolism of these meals? The debate has been between 
those who think that they did contain at least an element of Catholic 
'communion', such that the participants partook of the divinity when 
eating the meal, and those who reject any such interpretation and claim 
that sacred banquets in the mysteries were no different from any other 
post-sacrificial meals, or even funerary banquets, in antiquity 83 It seems 
to me, once again, that the dispute, at any rate in this form, cannot be 
resolved: there is so little direct information that a priori assumptions 
crowd in almost from the beginning, so that it is almost impossible to 



81 Invitations: Totti 1985, 125-27 lists 16 examples under her no. 48; also T.C. 
Skeat, Another Dinner-Invitation from Oxyrhynchus {RLond. inv. 3078, JEA 61 (1975) 
251-54 at 253 n. 2 (not all to the Sarapis4/me however). Add P.Coll. Toutie I. 51-52; 
POxji 4339 and 4540 (both with useful commentary); and see still H.C. Youtie, The 
'Kline' of Sarapis, HThR 41 (1948) 9-29 = idem, Scriptiunculae 1 (Amsterdam 1973) 
487-507 (on the interesting letter EMich inv. 4686 from Karanis, IIP = Totti 128f. 
no. 49, which contains the memorable phrase kcu yap avxtretv avGpawtoi; o\) Sijvaxat 
T(p rcupitp ZapdittSt, 1. 15f). Invitation to the 'Mine of Anubis' in a temple of Sarapis 
(late IIP), possibly a funerary meal: SB 14503 = D. Monserrat, The kline of Anubis, 
JEA 78 (1992) 301-07. Fire-brigade: Tertullian, Apolog. 39.15. This may not be quite 
as far-fetched as it seems: Ptolemaios in EMich inv. 4686 had to provide five ass-loads 
of wood for the feast. 

82 This has become something of a topos recently, cf. K. Dunbabin, Convivial Spaces: 
Dining and Entertainment in the Roman Villa, JEA 9 (1996) 66—80; eadem, The Eoman 
Banquet: Images of Conviviality (Cambridge 2003); the essays in Nielsen 1998; Egelhaaf- 
Gaiser 2000, 272—329. Defunctive commensality: M. Heinzelmann, Die Nekropolen von 
Ostia: Untersuchungen zu den Grdberstrqfien vor der Forta Eomana und an der Via Laurentina 
(Munich 2000); S. Stehmeier, Gemeinschaft uber den Tod hinaus. Grabtriklinien als 
Festplatze romische-kaiserzeitlicher Collegia, in Nielsen 2006, 215—24. 

85 Cybele seems to have been considered to be present at funerary banquets in 
Asia Minor, with the dead person in a sense being presented to her: E. Mitropoulou, 
The Goddess Cybele in Funerary Banquets and with an Equestrian Hero, in Lane 
1996, 135-165. 



THE RITUAL SYSTEMS 231 

adjudicate 'objectively'. 84 Since the cult-meal was a cultural reality, 
over time it gradually acquired more elaborate features, just as it did in 
Christianity We might say that the sacred banquet had both a pagan 
and a Christian form, and the decision as to which is to take precedence 
is largely a matter of the scholar's subjective choice. 

Seen in a functional light, the sacred banquet is important because 
of the different levels at which it can be seen to contribute to group 
cohesion. 85 Performatively it turns aspects of the mythical account of 
the natural cycle into action and direct speech, for greater immediacy 
of comprehension. Symbolically, it links the acceptance of death to the 
continuance of life, thus representing human life, and death, as a special 
case of a universal rhythm (cf. Rudhardt 1958, 60). The site where the 
ritual is celebrated is also of central importance, since it establishes links 
between locus, symposiast and deity. Mithraea are again an excellent 
illustration, since, at any rate ideally, they systematically reproduced 
the set of ideas through which the individual initiate could relate to 
the god's example, which in this case itself included a sacral banquet 
together with Helios/ Sol. None of this is peculiar to the mysteries, since 
comparable banquets are found all over the world (Bleeker 1963, 228), 
but, given the ideological biases of the comparative method, it seems 
preferable to examine each variant for itself. 

d. Prayer 

'hi, Paraetonium Mareoticaque arva Pharonque 
quae colis et septem digestum in cornua Nilum, 
fer, precor, ' inquit 'opem nostroque medere timori! 
te, dea, te quondam tuaque haec insignia vidi 
cunctaque cognovi, sonitum, comitesque facesque . . .' 

Ovid, Metam. 9.773-77 

Prayer is one of the least noticed, because most highly routinised, 
every-day activities. Though it may seem to be of minor importance, 



84 See also my remarks on the relation between Christianity and the mysteries in 
Chap. 5. 

85 Rudhardt 1958, 160; Bleeker 1963, 227. I would want to add that the ritual also 
expresses the internal hierarchy of the group: in performing his proper function, each 
member implicidy assents to the division of labour and the hierarchy of value associated 
with it, cf. J. Riipke, Organisationsmuster Rom, in Bonnet, Riipke and Scarpi 2006, 
13—26 at 24f (on the Torre Nova Bacchic inscription). However Rupke draws rather 
sweeping conclusions from a single inscription that makes no pretence of being more 
than a snapshot of the cult at a moment in time. 



232 CHAPTER FOUR 

it should in fact interest anyone trying to understand how ancient 
religion functioned (Versnel 1981a). Written texts are here crucial; for- 
tunately we dispose of a good deal of information, from literary and 
epigraphic sources, on ancient practice. We thus have an insight into 
the type of appeals made to the gods in a variety of situations, mainly 
of course requests, for blessing, for prosperity, health, for divine help 
and intervention as in the passage of Ovid cited in the epigraph, but 
also philosophical 'hymns', and a range of evocations, conjurations 
and curses. 86 

The Romans themselves evidently devoted some thought to the 
issue of prayer. Pliny the Elder, for example, provides a brief func- 
tional typology in the context of the question whether language has 
an inherent power (polleantne aliquid verba et incantamenta carminum). 81 He 
first mentions prayer in the context of sacrifice: sacrifice without prayer 
non videtur referre, just as divination (extispicy) without prayer seems to 
be valueless. He then distinguishes three sub-types of public or official 
utterance (obsecratio), which he rather clumsily terms impetritae, depulso- 
riae and commentation.es or commendationes (Pliny, //JV28.11). The first are 
invocations to obtain favour from the gods, in particular favourable 
omens, but perhaps also evocations and conjurations; the second are 
deprecationes, i.e. prayers to avert evil. The third category requires a little 
further discussion because the text seems to be corrupt. Most mss. read 
commentationis (V, R, d, v); one reads commentationes (E); commendationis is 



Sl> There are three good older works: C. Ausfeld, De Graecorum precationibus 
quaestiones, Jb. klass. Philol, Suppl. 28 (1903) 506-47 (also separately, Leipzig 1903); 
Appel 1909; PJ.T. Beckmann, Das Gebet bei Homer (Diss. Wurzburg 1932); more recently, 
note E. von Severus, s.v. Gebet, I, R/AC 8 (1972) 1134-1258; O. Michel and Th. 
Krauser, Gebet, II (Furbitte), ibid. 9 (1976) 1-36; Limet and Ries 1980; D. Aubriot- 
Sevin, Prieres et conceptions religieuses en Grece ancienne jusqu'd la fin du V siecle. Coll. Maison 
de l'Orient mediterraneen (Lyons 1992); EV Hickson, Roman Prayer Language: Livy and 
the Aeneid of Vergil (Stuttgart 1992); S. Pulleyn, Prayer in Greek Religion (Oxford 1997); 
Kiley 1997; C. Guittard, Invocations et structures theologiques dans la priere a Rome, 
REL 76 (1998) 71-92; W.D. Furley and J.M. Bremer, Greek Hymns. 2 vols (Stuttgart 
2001); J.C. Thom, Cleanthes' Hymn to Zjus: Text, Transl. and Comm. (Stuttgart 2005); 
G. Zuntz, Griechische Philosophische Hymnen (Tubingen 2005); E Graf, Prayer in Magic 
and Religious Ritual, in C.A. Faraone and D. Obbink (eds.), Magika Hiera (New York 
1991) 188—213. An anthology of early Christian prayer: A. Hamman, La priere dans 
VEglise ancienne (Berne 1989). 

87 C. Guittard, Pline et la classification des prieres dans la religion romaine (NH 
28.10—21), Helmantica 38 (1987) 157—80; idem, Recherches sur le carmen et la priere dans 
la Litterature latine et la religion romaine (Paris 1996). 



THE RITUAL SYSTEMS 233 

H. Barbaras' suggestion (1492). 88 Before 1972, most specialists in Roman 
religion opted for Barbaras' emendation as more appropriate to the 
idea of prayer (Koves-Zulauf 1972, 38 n. 35). In that year, however, 
T. Koves-Zulauf, in the most detailed examination of this passage ever 
undertaken, argued strongly on 'structural' grounds in favour of the 
mss. reading, interpreting it as a quite different type of official utterance, 
the (augural or haruspical) interpretation of a sign. 89 

In his review of Koves-Zulauf, Jerzy Linderski rightly argued, as 
Barbarus and many others had done before him, that a text of Valerius 
Maximus supports the conjecture commendationis by showing that the 
latter was indeed a type of precatio, as Pliny's text implies: 

Prisco etiam institute rebus divinis opera datur, cum aliquid commendandum est, 
precatione, cum exposcendum, voto, cum solvendum, gratulatione, cum inquirendum 
vel extis vel sortibus, impetrito, cum sollemni ritu peragendum, sacrificio, quo etiam 
ostentorum acfulgurum denuntiationes procurantur. 90 

Valerius Maximus, Mem. 1.1.1a, ed. Combes 

Since the previous sentence also alludes to the portentorum dupulsiones, 
the ritual aversion of portents, Valerius Maximus seems to have been 
familiar with the same principle of division as Pliny, even though he 
puts the elements in the order depulsiones-commendandum-impetritum. 91 
We may therefore take it that Pliny's third form of precatio was the 



88 L.Jan in 1858 suggested reading commendatoriis, but no one has followed him (cf. 
Mayhoff ed.min. ad loc). 

89 Koves-Zulauf 1972, 34-63. Of Koves-Zulauf 's reviewers, J. Andre, Latomus 33 
(1974) 195f. accepts the 'structuralist' method and commentatio without demur; R.M. 
Ogilvie, CR 26 (1976) 283f. seems to accept commentatio but disagrees about its signifi- 
cance; J. Poucet, AC 43 (1974) 483-86 accepts commentatio and praises the learning, but 
objects to the 'structuralist' method. Otherwise, the book seems to have evoked more 
dismay than interest: it was never reviewed in Gnomon, REL, RHR or JRS. 

9(1 "Ancient rules also regulated the conduct of ritual: prayer, when it was a matter 
of putting something into the gods' care; the votive undertaking, when it was a matter 
of asking for something and thanksgiving when it came to redeeming the vow; asking 
for good omens when it was a matter consulting the gods by means of the entrails or 
lots; sacrifice when it was a matter of regular worship — sacrifice was also used in cases 
when it was necessary to expiate portents and lightning-strikes." 

91 J. Linderski, CIPhil 70 (1975) 284—89. Linderski cites a number of parallel texts for 
the idea of 'putting something into the gods' care', e.g. Augustine, De civ. Dei 4.8 (the 
crops entrusted to Segetia); 4.2 1 (nurslings to Ops, and a whole list of others), Tacitus, 
Ann. 15.23 (Poppaea's unborn child to the gods); Apuleius, Met. 3.7 (a plight to the 
gods), etc. and plausibly links this religious usage to imperial commendatio to an office. 



234 CHAPTER FOUR 

commendation or entrusting of somebody or something to the care 
of the gods. 92 

It seems probable that there were other more complex Roman divi- 
sions of prayer-forms than this, which may for example have been 
discussed in the books de pontijicibus and de auguribus of Varro's De rerum 
divinarum, and perhaps also in the almost wholly obscure Book 1 1 {de 
consecrationibus). 93 From the surviving material, we certainly distinguish 
a wider range of utterances, or texts, that can loosely be thought of as 
prayers: invocations, supplications, hymns, vows and votives, aretalogies, 
oaths, a range of performatives such as purifications, 'judicial prayers' 
and regular imprecations or curses (Versnel 1981a, 17—26). However 
these three types (or four, including prayer at sacrifice) may be taken 
as those prayer-types that would most naturally have occurred to a 
Roman. 

Pliny's list reminds us of the obvious point that many private and 
public acts in the Roman world were ritualised, thus requiring divine 
approval. Such a close and repeated relation to the gods is obviously 
incompatible with the old idea (which one still comes across in New 
Testament circles, for example) that the population of the Empire was 
largely dissatisfied with their religion. Quite apart from the evidence 
for enthusiastic participation in civic festivals and sacrifice, there was 
a widespread individual religiosity that is everywhere evident in the 
literature, as well as the epigraphy, of the Empire. 94 Alongside this 
diffused religiosity, however, there was a very vocal level of philosophi- 
cal criticism of popular religious practice, represented for example by 
the Cynics, the Epicureans and the Academics. Scholars have in the 
past paid far too much attention to this discourse, simply because it 
is easily available in the form of argumentative texts. Historians have 
tended to ignore popular religious culture for another reason, being 
all too interested in the theological justification of empire, for which 
there is of course a great deal of fascinating material. 95 The traditional 



92 Pliny remarks on the emperor Tiberius' approval of offering good wishes when 
someone sneezed (28.23). 

95 E. Norden, Aus altromischen Priesterbiichern. Skrifter utgivna av kungl. humanistiska 
vetenskapssamfundet i Lund 29 (Lund and Leipzig 1939) 4—6. Cardauns lists just two 
frags, from de consecr., his nos. 83f 

94 Cf Versnel 1981a, 26-33; Veyne 1989; 1999. 

95 Cf. e.g. J. Rufus Fears, Princeps a diis electus: The Divine Election of the Emperor as a 
Political Concept at Rome. Monogr. of the American Academy at Rome 26 (Rome 1977); 
idem, The Cult of Jupiter and Roman Imperial Ideology, AMRWIl. 17.1 (1981) 3-141; 
idem, The Theology of Victory at Rome: Approaches and Problems, AJVRIY 11.17 .2 



THE RITUAL SYSTEMS 235 

Roman religious system, that is the sacra publica, supplemented by all 
that is implied by mos maiorum, being essentially the religion of the 
elite, did not need to be validated or legitimated by the mass of the 
population. It is now agreed that the (expanding) Roman elite, whether 
or not individuals had access to the philosophical critique of civic 
and popular religion, continued to subscribe to the performance of 
the post-Augustan sacra publica at Rome and, in a modified degree, in 
the coloniae; the aim of Roman religious 'tolerance' was to leave the 
religious practice of all constituent cities and peoples intact so far as 
possible, since there was no means, except for the institutions of the 
imperial cult, of linking centre with periphery in this regard. Insofar as 
integration between the systems occurred, it did so through the initia- 
tive of local elites, especially within the municipia and a'yz'to-capitals of 
the Latin-speaking west. 96 

I do not think that the impulse behind the expansion of the oriental 
cults in the Hellenistic-Roman world had anything to do with dissatis- 
faction with an existing system, let alone its radical failure. Rather their 
success was due to a variety of factors: the demand in an increasingly 
differentiated world for differentiated religious experience beyond the 
confines of the family; the provision of relatively specialised religious 
experiences in a context in which religion was increasingly seen as a spe- 
cial order of discourse and practice; the creation of relatively accessible 
yet specialist social contexts where knowledge of a new and significant 
kind could be exchanged; the offer of a variety of new symbolic roles 
and functions. Among these factors, prayer, or, to put it more generally, 
specialised discursive opportunities, should not be underestimated. 

We may take it those who got to the point of being initiated into one 
or other of the oriental cults considered themselves, and were considered 
by most others in their social groups, as pious (by piety in this context I 
mean the strict observance of the demands made by the gods). 97 Indeed, 
it seems very likely that the social and religious integration of initiates 



(1981) 736-826; idem, The Cult of Virtues and Roman Imperial Ideology, ANRW 
II. 17.2 (1981) 827— 948; J.-P. Martin, Providentia Deorum: Recherches sur certains aspects 
religieiux du pouvoir imperial remain. CEFR 61 (Rome 1982); M. Bergmann, Die Strahlen 
der Herrscher. Theomorphes Herrscherbild und politische Symbolik im Hellenismus und der romischen 
Kaiserzeit (Mainz 1998). 

96 T. Derks, Gods, Temples and Ritual Practice: The Transformation of Religious Ideas and 
Values in Roman Gaul. Amsterdam Archaeological Studies 2 (Amsterdam 1998); Spicker- 
mann 2003. 

97 Cf. H. Wagenvoort, Pietas, in idem, Pietas. Selected Studies in Roman Religion (Leyden 
1980) 1-20. 



236 CHAPTER FOUR 

into these cults went hand in hand. The typical expression of this piety 
is a self-consciously humble order of discourse, submissive to the will 
of the increasingly grand, indeed almighty, deity 98 The ostensible aim 
is not to achieve the concrete end one wants; rather the deity is to be 
moved by the sincerity of the appeal made to him or her, such that the 
individual can share in the formation of his own destiny. The claim to 
be a servant of the deity (vnr\peTr\q, SovXoc,, Xdctpiq) is taken as proof 
of pietas. In Latin, famulus, slave, of a god becomes a perfectly accept- 
able self-description: after his first initiation, Lucius comes to realise 
that he is being asked to become a. famulus of Osiris too: prohinc me 
quoque peti magno etiam deo famulum sentire deberem (Apuleius, Met. 11.27). 
The word seems to have been borrowed from a locution, perhaps 
even a sort of technical term, in the cult of Cybele, most familiar to 
us in a negative sense, as in the lament of Catullus' Attis: Ego nunc eum 
ministra et Cybeles famula ferar? (1. 68, cf. 90; 52)." After two centuries of 
Christian use of the word to express a humble relationship to God, by 
the late fourth century we even find Aconia Fabia Paulina, the wife of 
the extremely aristocratic Vettius Agorius Praetextatus, calling herself, 
or rather announcing herself as being described by her husband, as 
famula divis. 100 

We may perhaps invoke a late fourth-century Christian source to give 
us an idea of the emotional attitudes implied by the mysteries. Evagrius 
of Pontus (c. 346-99) wrote a workD<? oratione (rcepl npoazvxi\<i), in which 
he gives us a good idea of the attitudes towards prayer in his own day, 
irrespective of the particular brand of worship. 101 "Prayer," he tells 



98 Nock 1925; Bomer 1957-63, 3: 207f.; Pleket 1981, 159-71; D.B. Martin, Slavery 
as Salvation: The Metaphor of Slavery in Pauline Christianity (New Haven 1990); Alvar 1999; 
C. Hezser, Jewish Slavery in Antiquity (Oxford 2005) 348: "The listener/reader is invited 
to identify him- or herself with the slave and/or the other subordinate characters" (on 
Midrashic king/slave parables). 

99 Other negative occurrences in relation to Cybele: Cicero, De leg. 2.22; Livy 37.9.9; 
Valerius Flaccus 3.20; Pentheus: Ovid, Met. 3, 574; of an actor/delator: Martial 9.28.10 
(famulus Iovis). Occasionally in a positive sense: Horace, Epist. ad Pisones (Ars poetica) 239 
(Silenus custos ac famulus of Bacchus); Vergil, Aen. 1 1 .558 (of Camilla); Germanicus, Aral 
38 (of Cybele); Manilius 4.760 (of Cybele's lions); Seneca,^. 255 (of Cassandra). 

100 CIL VI 1779, rev. 1. 24 = ILS 1259 = CCCA 3: 62-4 no. 246. 

101 Text ap. I. Hausherr, Le kgons d'un contemplatif Le traite de I'Oraison d'Evagre le Pon- 
tique (Paris 1960); tr. M.-O Goudet, ap. Evagre: de la priere a la perfection (Paris 1992); cf. 
A. Guillaumont, Un philosophe au desert: Evagre le Pontique, RHR 181 (1972) 29-56; 
idem, Aux origines du monachisme chretien. Pour une phenomenologie du monachisme (Bellefontaine 
1979); E.A. Clark, The Origenist Controversy. The Cultural Construction of an Early Christian 
Debate (Princeton 1992) 66-84. 



THE RITUAL SYSTEMS 237 

us, "is an intimate spiritual relation to God" (§3). A little later there 
appears an assumption about 'true' prayer that seems to be character- 
istic of the entire late Roman period: "Make sure all your prayers are 
accompanied by tears. They guarantee success. The Lord delights in 
tearful prayer" (§6). We may note a further point surely valid for the 
'oriental' mysteries too: "Do not rely solely on external gestures and 
comportment: direct your mind towards the idea of fervent spiritual 
prayer" (§28). Such injunctions are the behavioural correlate of a self- 
definition as 'servant'. 

Not all prayer was of this type however. The oriental gods were 
nothing if not \)7ir|Kooi, etctikooi, dipraesentes. The motives for invocation 
were naturally very varied, though many will have been prompted by 
extraordinary circumstances or immediate needs. A series of Hellenistic 
epigrams, the earliest dating from the later third century BC, thematise 
dedications by galli in thanks for divine intervention: in one ascribed to 
Dioscorides, Atys, on the road between Pessinus and Sardis, takes shelter 
as night falls in a cave. A savage lion appears; xwoq ocopr| Soc{|aovo<;, 
Atys beats the tympanon, thus frightening the beast away; in gratitude to 
her, he dedicates either the cave itself or a portable model of a shrine 
(9aA,duT|) and the tambourine to the goddess. 102 Despite the detailed 
variation, some of it apparently comic, all emphasise Cybele's miracu- 
lous saving intervention. The miracle of Claudia Quinta fits perfectly 
into this tradition: mota dea est sequiturque ducem laudatque sequendo (Ovid, 
Fasti 4.327); Kcxi r\ Qebq eo7iexo (Appian, Hann, 7.56, 234.3). 103 For her 



102 Dioscorides ap. AnthPal. 6.220 Waltz = Gow-Page, 11. 1539-54, Dioscorides no. 
16 with their comm.; the others are Alcaeus ap. AnthPal 6.217 = Gow-Page, 11. 134—43, 
Alcaeus no. 21; Antipater, AnthPal. 6.219 Waltz = Gow-Page 11. 608-31, Antipater 
no. 64; Simonides ap. AnthPal 6.217 Waltz = Gow-Page, 11. 3504-3313, 'Simonides' 
no. 2; on the motif, cf. D. Gall, Catulls Attis — Gedicht im Licht der Quellen, WJA 
23 (1999) 83-99; Harder 2004, 577-79. As Gall points out, Catullus seems to have 
inverted this topos in introducing the chastening lion at 63.75—90. For the portable 
naiskos rather than an entire cave, see A.S.E Gow, The Gallus and the Lion, JHS 80 
(1960) 88-93 at 93. 

103 Bomer ad loc. IV 300 notes the typological quality of the narrative of the casta 
persona who frees the way for the impeded god/saint to proceed; cf. J.N. Bremmer, Slow 
Cybele's Arrival, in idem and N.M. Horsfall (eds.), Roman Myth and Mythography. BICS 
Suppl. 52 (London 1987) 105—11. The 'commemorative' statue of Claudia Quinta 
in the vestibule of the Palatine temple of Mater Magna was itself later spared by no 
less than two fires, one in 1 1 1 BC, the other apparendy in the late Republic; at any 
rate, Augustus certainly rebuilt it in 3 BC (Val. Max. 1.8.1 1; Tacitus, Ann. 4.64.4 with 
Furneaux ad loc; Augustus, RG §19 Malcovati; cf. P. Pensabene, s.v. Magna Mater, 
Aedes, LTUR 3, 206f). 



238 CHAPTER FOUR 

part, Isis has a good claim to be called f| ocoxeipa par excellence. 104 
One of Isidorus' hymns at Medinet Madi claims: 

oacoi 8' e|i |ioipai<; Gavdtov cruvexovtai ev eipK-rpi, 
Kai oaoi dypimviatf; |ieydXai<; oxA-cuvx' 68wripa'i<;, 
Kai oi ev dM.OTpir|i xSpT^ 7i^.avo(0|ievoi dv8pe<; . . . 
^co^ovS' ootoi anavxeq, e7ie\)^d|ievoi oe itapeivai. 105 

Hymn I, 11. 29-34 (Bernand 1969, 633) = Vanderlip 1972, 17f. 

The most expressive word here is napzivav. what is desired is the actual 
presence of the deity. Indeed, the narrative recording of miracles vouch- 
safed in response to prayer, as in the epigraph, was a regular sub-genre 
in the cult of Isis and Serapis. 106 The transformation of Lucius back 
into a human-being is an over-familiar example. 10 ' A workshop, perhaps 
in Rome, during the second century AD produced a minor genre of 
amulets hung round the neck to protect children from dangers that carry 
the epiklesis veikoc f| Eton;. 108 Given that the main Isiac festival in the 
Graeco-Roman world (which seems not to have been Egyptian) 109 was 
the Isidis navigium on 5th March, it is hardly surprising that much of 
our evidence for Isis' intervention relates to sea-farina;. In the standard 



101 Ecoxeipa is one of her most frequent votive epithets: SB 10. 10683 (Alexandria, 
III-IP); iS/OS 202/0 170 (Sarapieion C, Delos; with Sarapis), 202/0365 (ibid.), 204/1004 
(Kos). Originally the concept seems to have applied solely to Ptolemy IV Philopator 
and his wife Arsinoe (cf. RICIS 402/0601 [dedicated after the battle of Raphia in 
217 BC] and the texts there cited, but was evidently soon 'democratised'. Even more 
common for Isis is iitT\Kooq (RICIS lists 13 cases). Among many concrete examples, 
note Protos son of Pythion otoGeli; ek koXX&v Kai ireydXcov KivSwcov (Sarapieion C, 
Delos, mid-IP): RICIS 204/0108; P. Aurelius Dioskouros at Lepcis Magna, ek \izyaXi\c, 
vooot) 8iaoco9ei<;: AE 2003: 1903b 1. 14f. (to Sarapis and cruvvaoi 9eoi); a librarius of 
the Colonia Ostiensis, P. Cornelius Victorinus, dedicated a statuette of Mars with a 
horse to Isis Regina, restitutrici salutis meae: CII XIV 343 = RICIS 503/1 1 18. 

103 "As many as are bound fast in prison, in the power of death, As many as are 
in pain through long, anguished, sleepless nights, All who are wanderers in a foreign 
land . . . All (these) are saved if they pray that You may be present to help", tr. Vanderlip 
(= Totti no. 21). 

106 Cf. Totti nos. 12-17. 

107 On the transformation as a typical marvel, cf. Egelhaaf-Gaiser 2000, 80—84. 

108 G. Sacco, Un amuleto isiaco dalla via Latina, in Epigraphica: Atti delle Giornate di 
Studio di Roma eAtene in memoria di Margherita Guarducci (1902-1999). Opuscula epigraphica 
10 (Rome 2003) 141-50 counts seven cases. 

109 Cf. Bergman 1968, 198—205; P. Bruneau, Isis Pelagia a Delos (complements), 
BCH 87 (1963) 301-08 at 306-08; Bricault 2006b, 16 and 177. J.G. Griffiths, The 
Egyptian Antecedents of the Isidis Navigium, Studia Aegyptiaca I. Recueil d'etudes dediees 
a Vilmos Wessetzky a I'occasion de son 65' anniversaire (Budapest 1974) 129-136, argued that 
there were Egyptian precursors. 



THE RITUAL SYSTEMS 239 

self-predication, she claims to have invented QaXaaaia epycx, matters 
nautical. 110 In the 'literary' version from Andros, she it was who first hit 
upon the idea of praising men's nautical endeavours. 111 At Pelusium, 
on the most easterly arm of the Nile Delta, she was named opuiaxpioc, 
the one who brings the voyager safely back to harbour {P.Oxy. 1380 
1. 73f.). 112 Pausanias mentions that in his day there was a shrine to Isis 
Pelagia just off the agora of Corinth on the road up to Acrocorinth. 113 
All over the eastern Mediterranean thanks were offered her for the safe 
accomplishment of a voyage, attested by (rather uncommon) votives to 
Isis Euploia, Pelagia or Pharia, lu or perhaps to Isis Fortuna, who is often 
shown with a rudder. 115 One of the earliest accessions to the British 
Museum was a fine pottery lamp found in the sea off Pozzuoli, in the 
shape of ship being steered by Isis and Serapis and bearing the epiklesis 
EvnXom, "Safe voyage!". 116 In all these cases, ocotripioc or salus is to be 



110 RICIS 302/0204 1. 15 (Kyme, also Ios [202/1 101] and Thessalonike [1 13/0545]); 
also the charts in Bruneau 1974, 353—55; P. Bruneau, Isis Pelagia a Delos, BCH 85 
(1961) 435-46 at 443; Vanderlip 1972, 33f. 

111 IG XII.5, 739 = Totti no. 2 = RICIS 202/1801 1. 34f. (F); also the poetic embel- 
lishment in 11. 145-57. 

112 For the location, see Barrington, Map 72, H2. The epithet is not attested epi- 
graphically. 

113 Paus. 2.4.6 with Musti/Torelli 2 p. 231f. ad loc; cf. D.G. Smith, The Egyptian 
Cults at Corinth, HThR 70 (1977) 201-31; E.J. Milleker, Three Heads of Sarapis from 
Corinth, Hespena 51 (1982) 121-35. Bruneau 1974, 337 pointed out that SIRIS 34, to 
ETcuSi neXayia euaKOG), supposedly found at Corinth but now lost, is afalsa copied from 
RICIS 205/0302 (Mytilene, also lost). It is duly omitted from RICIS by L. Bricault. 

111 The sole votive examples of Euploia are: RICIS 202/0329 (Sarapieion C, Delos, 
?IF); 202/0365 (ibid., ?II-F); of Pelagia: 205/0302 (Mytilene, ?F; see previous note); 
305/1402 (Iasos, Roman period; reads "Io"i[Si lie X]ayia); of Pharia: 402/0501 (Balaneia 
in Syria-Phoenicia, Roman); 503/1204 (Portus, ?II P ); see now the discussion in Bricault 
2006b, making the same point. For a sample of such images, see Tran tam Tinh 1990, 
1: 782-86 nos. 269-315. Bruneau 1974 objects to the denomination 'Isis Pelagia' for 
the (mainly numismatic) image of Isis standing on a ship, with her tunic or veil bellying 
out as a sail, preferring 'Isis a la voile'. Bricault 2006a, 84 agrees. 

115 The suggestion that Isis Fortuna often conceals Isis 'Pelagia' was made by Mal- 
aise 1972a, 180f, followed by Bruneau 1974, 379-81. Bricault suggests that awxeipa 
in the Aegean islands had the same force (ap. RICIS 204/0108); cf. the graffito at 
the entrance to the harbour at Pompeii, Eiauuxri acp^ovca: CII IV 4138 = RICIS 
504/0216. Bricault 2006a, 84—90 -90 assembles a mass of numismatic and dactylic 
evidence for the maritime engagement of the Egyptian deities, showing how closely 
safe sea-journeys were linked to the wider ideas of prosperity and 'salvation'; see also, 
in greater detail, idem 2006b. 

116 H.B. Walters, Catalogue of lamps in the British Museum (London 1914) no. 390 with 
pi. X = IG XIV 2045.48 - Merkelbach 1995, pi. 213 = RICIS 504/0403 (I-IF). The 
inventory no. is 100. 



240 CHAPTER FOUR 

understood concretely, just as in the case of the numerous votives set 
up in gratitude for being healed or cured. 117 

3. Rituals in the Phrygian Cults 

a. Introduction 

The transfer of Cybele to Rome had implications both for the cult's 
ritual dimension as well as for its institutions, real and imaginary. 118 
According to Livy the image that was brought to Rome in 204 BC 
and placed in the temple of Victoria on the Palatine was a plain stone, 
in which the deity was believed to dwell. 119 Late-Roman images of the 
stone as it was borne into the Circus Maximus on aferculum for the 
Megalensia suggest that the object then in use was a squat, roughly conical 



117 E.g. P. Aurelius Dioskouros at Lepcis Magna, etc \ytyaXr\c, voaoi) SiaacoBeic;: AE 
2003: 1903b 1. 14f. (to Sarapis and owvotoi Geoi). 

118 A more or less complete list of the sources will be found in Erskine 2001, 206. 
The general credibility of the major narratives of Livy 29.10.4—11.8; 14.5—14 and 
Ovid, Fasti 4.247-348 has been impugned by e.g. Schmidt 1909, 1-18; Bomer 1964 
(cf. his comm. on the Ovid passage, 2: 228—30) and more recently Berneder 2004. 
Even if we agree with Wissowa that Schmidt "freilich in radikaler Skepsis zu weit geht" 
(1912, 318) it is far from clear which details are reliable and which fiction. It is really 
a matter of deciding whether, with the majority, to support Livy, cf. most recently PJ. 
Burton, The Summoning of the Magna Mater to Rome (205 BC), Historia 45 (1996) 
36-63; Erskine 2001, 212-18; or, with Gruen 1990, 5-33, Ovid, esp. Fasti 4.331-48, 
cf. 181—87. I take the latter view: J. Alvar, Esconografia para una reception divina: la 
introduction de Cibeles en Roma, DHA 20 (1994) 149-69. 

119 Livy 29.1 1.7; 14.14. Livy uses the expression lapis sacer; the word for such objects 
in English, Germanic and Romance languages is some version of Greek [3outoA,o<; 
(etym. unknown), transliterated into Latin as baetulus; Pliny, HM 37.135 quotes from 
the early IV a Greek lithographic writer Sotakos the information that there are two 
sorts of gemmae cerauniae, semi-precious stones that fall from heaven; of these, baetuli 
are roundish, black ones, which are sacred, look like axe-heads, and could be used to 
capture cities. Damascius, Vit. Isidori §94 p. 138 Zintzen claims that they were common 
on Mt Lebanon near Heliopolis; elsewhere he describes their marvellous properties, as 
a manifestation of the god Gennaios = Genneas (§203). However, no ancient source 
authorises the use of this word for the sacred stones of Cybele (of which there were 
many); cf. C. Auffarth, s.v. Baitylia, I, BMP 2 (1997) 405; W.K. Pritchett, Pausamas Per- 
iegetes. 'Apxccta "EXXai; 6 (Amsterdam 1998) 99-101. For the ayaX)j,a falling from the 
sky: Herodian 1.11.1; although, as is clear from Pausanias' usage, the word need not 
mean a statue in the usual sense, Herodian clearly assumes it was; hence his complaint 
that the name of the sculptor is unknown. 



THE RITUAL SYSTEMS 24 1 

stone with a flattened top. 120 It thus had no marked, or even recognisable, 
phallic symbolism. Of course by 204 BC the Mater Magna had already 
for at least two centuries been represented in Greek anthropomorphic 
form: the Romans knew her cult from Athens, where there had been 
a public temple to her in the Agora, the Metroon, with its famous 
'Pheidian' statue of the goddess, since at least the late fifth century; 121 
from the Piraeus, where the private cult, documented already in 281/0, 
was transformed into a semi-public one (conducted by the orgeones, i.e. 
without any 'oriental' features) in the late third century, or even earlier; 122 
from Delos, where there was a Metroon at least as early as 208 BC (as 
at Athens, it served as a public archive), and where several images of 
the goddess are listed in temple inventories both before and after 166 
BC; 123 and finally from Sicily, for example at Colle Orbo and the major 
site at Akrai, which was clearly modelled on Anatolian shrines, such as 
that of Kapikaye at Pergamum. 124 At the end of the Hannibalic war, 
P. Scipio Africanus famously dedicated some spoils at her shrine in the 
city of Engyum, later raided by Verres (Cicero, 2Verr. 4.97; 5.186). 125 The 
claim to have acquired a stone — certainly not 'the' stone but perhaps 



120 Both forms, Megalensia and (ludi) Megalenses have good authority; I follow the form 
usual in the late Republic and early Principate. 

121 H.A. Thompson and R.E. Wycherley, The Athenian Agora, 14: The Agora of Athens: 
The History, Shape and uses of an Ancient City Center (Princeton 1972) 29—38, arguing still 
for a high date, as Thompson had done from 1937; J.S. Boersma, Athenian Building 
Policy from 561/0 to 405/4 BC (Groningen 1970) 31-34; R.C.T Parker, Athenian Religion: 
A History (Oxford 1996) 159f, 188-98; M. Munn, The Mother of the Gods, Athens, and 
the Tyranny of Asia (Berkeley and Los Angeles 2006) 58-64, 317-49, all for different 
reasons stressing the late-V a interest in the goddess, for a low one. Even if, as Noel 
Robertson argues, The Ancient Mother of the Gods. A Missing Chapter in the His- 
tory of Greek Religion, in Lane 1996, 239—304, Mritrip 9ewv was an indigenous deity, 
there can be no doubt, from the archaeological evidence found in the Agora, that 
she was assimilated to Cybele by the mid/late-fourth cent. BCE at the latest (cf IG 
II III 2 1.2 no. 1257, 324/3 BC) and probably by the late V a . For Agorakritos' statue, 
note Pausanias 1.3.5 with Beschi and Musti ad loc. (p. 272 1. 52), Pliny, HN 36.17; 
Arrian, Periplous 9. 1 (A. Silberman's commentary in the Bude p. 30 adds nothing); cf. 
Naumann 1983, 159-69. 

122 IG II III 2 1.2 no. 1273 (281/0 BC, still a thiasos); 1314-16 (213/2, 211/0 and 
'late III"'); later texts: 1327-29; cf. Nilsson, GGR 2: 114f; Vermaseren 1977, 35; 
Borgeaud 1996, 31-55. 

12! Bruneau 1970, 431—35. The Metroon may have been situated at the end of the 
dromos of Sarapieion C, built in the late third cent. BC. 

121 Cf. Sfameni Gasparro 1973, 114—55; eadem, Per la storia del culto di Cibele 
in Occidente, il santuario rupestre di Akrai, in Lane 1996, 51—86; eadem 1999, 
359-64. 

123 Diod. Sic. 4.9—80 however claims that the temple was dedicated to the Cretan 
Mf|Tepe<;. 



242 CHAPTER FOUR 

one of very many stones associated with the goddess — from 'Pessinus' 
through the good offices of Attalus I Soter (241-197 BC) must therefore 
be interpreted as a sort of one-upmanship, since in later Rome there 
was no dearth of perfectly normal representations of the Mater Magna 
on her throne (see Frontispiece). 

The standard accounts, ancient and modern, of the early history 
of the Phrygian cults at Rome insist on the fact that it was not simply 
a new deity that was introduced but an entire cult, complete with a 
liturgy, processions and cult-attendants from a religious tradition quite 
different from the Roman (cf. Ferri 2006, 236—8). As Mommsen, true 
son of a north-German Lutheran pastor, put it in a passage of memo- 
rable distaste: 

So muflte dennoch der wiiste Apparat der "groflen Mutter", diese mit 
dem Obereunuchen an der Spitze unter fremdlandischer Musik von 
Pfeifen und Pauken in orientalischer Kleiderpracht durch die Gassen 
aufziehende und von Haus zu Haus bettelnde Priesterschaft und das 
ganze sinnlich-monchische Treiben von wesentlichsten EinfluB auf die 
Stimmung und Anschauung des Volkes sein. 126 

On the authority of a well-known passage of Dionysius of Halicarnas- 
sus, it is generally thought that some of these features were at once 
found so foreign and dreadful that they had to be forbidden by a decree 
of the Senate: 

The praetors perform sacrifices and celebrate games in her honour every 
year according to the Roman customs (kcitcx xovc, T(0|iai(ov v6|aoi)<;), but 
the priest and priestess of the goddess are Phrygians, and it is they who 
carry her image in procession through the city, begging alms in her name 
according to their custom (\rr\Tpayvpxox>VTE<;, cocntep avxoic, e0o<;) and 
wearing figures upon their breasts (tvkovc, 7iepiKei|ievoi toi<; crrr)0£cn), 127 
and striking their timbrels while their followers play tunes upon their 
flutes in honour of the Mother of the Gods. But by a law and decree of 
the Senate, no native Roman (Tco|iai(ov tcov ocuSiyevcov oike . . .) walks in 
procession through the city arrayed in parti-coloured robe, begging alms 
or escorted by flute-players, or worships the goddess with the Phrygian 
ceremonies (oike opytd^et xr\v 0e6v toi<; <ppvyioiq 6pyi<xo|ioT<;). 

Dionysius of Halicarnassus, Ant. Rom. 2.19.4E (tr. Spelman-Cary) 



126 Mommsen 1889, 1: 869f. 

127 With reference to an incident in Galatia in 189 BC, Polybius uses the expression 
e'xovcei; rcpO0TT|9i8ia kcu twcoin; (21.37.6, cf. 6.7 with Walbank ad loc). These objects 
clearly belonged to the full regalia of such priests. 



THE RITUAL SYSTEMS 243 

The literary tradition thus offers a sharp contrast between the account 
of the triumphant, indeed miraculous, entry in 204 BC, with a happy 
people bringing gifts to the goddess and a rapid curtailment, even pro- 
scription, in the following years. 128 The explanation offered in antiquity, 
and largely followed today, was designed to mask this contradiction. 129 
The first celebration of the Megalensia, apparently already in 204 BC, 
later, from 194 BC, with ludi scaenici (Livy 34.54.3), and the dedication 
of the temple of the Mater Magna on the Palatine with ludi scaenici on 
10th April 191 (Livy 36.36.3f. with Briscoe ad loc), naturally involved 
the imposition of a regular Roman festival upon the strange new cult. 130 
These arrangements were duly carried out by the praetors, as we have 
seen, yet the priests of the cult, all of them foreigners, impeded the 
attempts at Romanisation by insisting on maintaining their traditional 
rites. "Ce n'etait pas moins a une Orientale, inconnue aux milieux 
italiques, que l'aristocratie romaine assurait son appui et que la foule 
portait ses dons, au sanctuaire de la Victoire, son hotesse provisoire" 
(Bayet 1957, 152). 

As it stands, this account contains two main problems. Given the 
familiarity of the Roman elite with the Hellenised East, 131 it seems 
highly unlikely that the Senate was so ignorant of the real nature of 
the Phrygian cults that it went ahead and introduced them on the say- 
so of the Sibylline oracle, and then found it had to curtail the religious 
practices actually involved. 132 Even if we assume that the Senate was 
working on the supposition that it was only admitting an 'authentic' 



128 Wiseman 1984; Beard 1994; Nauta 2004, 618-25. 

129 E.g. Latte 1960, 259£: "Es ist nicht wahrscheinlich, daB der Senat von dem Wesen 
dieses Kultes eine Hare Vorstellung gehabt hat"; cf. E Altheim, Rb'mische Religionsgeschichte 
(Baden-Baden 1953) 2: 51; Vermaseren 1977, 39; Borgeaud 1995, 101-3; Beard, North, 
Price 1998, 1: 97; Roller 1999, 264-71. 

130 For the temple, which had a massive podium (78 x 113 R feet = 17.10 x 
33.18m), see Pensabene 1978; 1988; Richardson 1992, 242; Stamper 2005, 118 with 
p. 117 fig. 88. 

131 Cf. G. Clemente, Esperti, ambasciatori del Senato e la formazione della politica 
estera romana tra il III e il II secolo, Athenaeum 54 (1976) 319—52; E. Gabba, Aspetti 
culturali dell' imperialismo romano, Athenaeum 65 (1977) 49—74; E.S. Gruen, The Hel- 
lenistic World and the Coming of Rome (Berkeley and Los Angeles 1984) 1: 203fE, 316ff 
(mutual knowledge); cf. idem 1990, 1—33. The same point is made from a different 
perspective by P. Oliva, Die Wolken im Westen. Griechenland und die Ankunft der 
Romer, Gymnasium 100 (1993) 1—18; on the later period, cf. J. Ferrary, Philhellenisme et 
Imperialisme. Aspects ideologiques de la conquete romaine du monde Hellenistique. BEFAR 271 
(Rome 1988). 

152 Sfameni Gasparro 1999, 374ff and M. Monaca, La Sibilla a Roma: I libri sibillini 
fra religione e politica. Collana di studi storico-religiosi 8 (Cosenza 2005) 216—20 revive 



244 CHAPTER FOUR 

version of the hellenised cult, such ingenuousness hardly fits with its 
usual well-informed caution. If the cult was later repressed or reformed, 
the reasons must have been different. Moreover, had the outlandishness 
of the cult's rituals really been as marked as the literary tradition claims, 
it would have been quite unnecessary to repress it, since no Roman 
would have been attracted to it in the first place. 

I therefore incline to think that the true reasons for the repression of 
the cult (assuming it to be a fact about the early second century BC) had 
nothing to do with the exoticism of its rituals. If we have to go along 
with the literary narratives, I prefer a political account. In my view, the 
literary account of the reception of the goddess, which highlights the 
role of the old aristocracy, was a means of masking antecedent popu- 
lar unrest expressed in a religious idiom. This popular pressure, which 
evidently included demands that the Mater Magna be brought to Rome 
in order to guarantee victory over Hannibal, was simply occluded by 
the version that made the dominant class responsible for her introduc- 
tion. This version attempted to represent the goddess' introduction as 
legitimate in a threefold sense: inspired by consultation of the Sibylline 
books, approved by the Senate, and centred on the figure of Scipio 
Nasica. A scenario such as this would account for the claim that the 
Senate approved the entry of the cult even though it must have known 
perfectly well what it involved; and explain why, having failed to win 
the first round, it soon set about 'domesticating' it by assimilating it to 
its own ritual models, forbidding direct participation in the processions, 
and forcing the people into the role of passive spectators. 133 

The attraction of this view is that it understands the Senate's (major- 
ity) position as consistently hostile to the introduction of the Phrygian 
cult; the 'Roman order' was always in fact the order of the dominant 
class. This class was worried by the implications for its own religious and 
existential rationality of the cult's orgiastic character and the practice 
of self-gelding Even the preliminary symbolic effort at Romanisation, 
the lavatio of the stone in the Almo, a tributary of the Tiber, could 
not keep such contamination at bay. There followed repression or 
partial occlusion, itself reluctant testimony to the fear that the assimi- 



the old suggestion that the Xviri were following the model of the introduction of Venus 
Erycina in 217 BC. 

133 Scullard 1973 notoriously managed to discuss this period in detail without ever 
referring either to the Mater Magna or the Megalensia. Religion is religion and politics 
is politics. 



THE RITUAL SYSTEMS 245 

lation had proceeded far more rapidly than the dominant order could 
wish. We might find support for such a reading in Valerius Maximus' 
account of the people's hostile reaction to the decision of the aediles 
for 194 BC, C. Attilius Serranus and L. Scribonius Libo, allegedly on 
the recommendation of the consul Scipio Africanus, to separate the 
senatorial seats at these first ludi scaenici at the Megalensia from those 
of the people — the beginning of the famous fourteen rows. 134 Rather 
than read this simply in terms of hellenisation, I find it tempting to 
see it as part of an ongoing conflict over the acceptability of the cult 
of the Mater Magna. Moreover, it has recently been stressed that, in 
view of the number of later references to her, there must have been a 
powerful (popular) oral tradition relating to Claudia Quinta and her 
marvel (Erskine 2001, 218). This could be read as popular resistance 
to the official version. 

This political argument however does have two main weaknesses. 
In the first place, the high aristocracy used the festival of the Mater 
Magna, according to Cicero in the year that Cato the Elder was 
quaestor, i.e. already in 204 BC, as the occasion for lavish banquets, the 
mutitationes. 133 The argument of general hostility to the goddess cannot 
account for this rapid assimilation of the cult into the socio-cultural 
modes of the aristocracy 136 The second is that it has no place for the 
one archaeological fact that we know about the early popular cult of 
the Mater Magna at Rome, namely that, as many of the terracotta 
votives found in the area of the Palatine temple by Guido Boni and 
later by Pietro Romanelli in the 1950s indicate, it was focused largely 
on the child Attis, already in the second century BC. 13/ To judge from 



134 Valerius Maximus 2.4.3, cf. 4.5.1. The tradition is complex: for the texts, cf. 
T.P. Wiseman, Phoenix 27 (1973) 189-98 at 195; Livy 34.44.5, with Briscoe ad loc, 
suppresses the detail about the people's hostility; cf. R.C. Beachem, The Roman Theatre 
and its Audience (London 1991) 62f. 

135 Cicero, De senect. 45 with J.G.E Powell ad loc; Fasti Praenest. prid. Non. April. (CIL F 
p. 235): Ludi M.D.M.I. Megalensia vocantur quod ea dea Megale appellatur. Mobilium mutitationes 
cenarum solitae sunt frequenter fieri quod Mater Magna ex libris Sibullinis arcessita locum mutavit ex 
Phrygia Romam; Aulus Gellius, NA 18.2.1 1; cf. Preller 1858, 451; Wissowa 1912, 318f; 
Becher 1991, 163; Rupke 2007a, 144. 

136 Cf. Ovid, Fash 4.353-56 with Bomer ad loc; Aulus Gellius, NA 2.24.2 reports a 
SC of 161 BC evidently shortly before the Lex Fannia of the same year, specifically 
limiting expenditure and display at these mutitationes; cf. J. Andre, L 'alimentation et la 
cuisine a Rome 1 (Paris 1981) 224; missed by Scullard 1973, 222. 

137 Romanelli 1963; a useful check-list in CCCA 3: 11-36 nos. 14-199, pis. XXI 
XCVII (not all Attides however). Romanelli considered them to date from c. 25 BC, 
but in fact they date mainly from the II P , being buried in the favissae after the fire of 



246 CHAPTER FOUR 

this important but still neglected material, the population of Rome in 
the second century BC saw the Phrygian Great Mother, as evidently 
already much earlier in Greece, as one of several female deities that 
offered succour, protection and healing to family and children. 138 Attis 
was interpreted simply as a representative child. The cult was indeed 
popular, but in a manner completely different from the impression of 
unbridled, un-Roman excess given by Ovid and Dionysius of Halicar- 
nassus. Archaeologically there is simply no evidence for the exoticism 
that looms so large there. If only we could be sure that that were the 
whole story, the distortions of the historical record might indeed be 
taken to be almost as far-reaching as Ernst Schmidt claimed in 1909. 

b. Emasculation 

Semper fac puer esse veils 

Ovid, Fasti 4.226 

ek |j,ev Y«P av8pcov yaXkoi yiyvovtat, £K 8e yd^iAcov 
av8pe<; ox> yiyvovxoa 

Arcesilaus of Pitane (attrib.) ap. Diogenes 
Laertius, Vit. phil. 4.43, p. 287.17-19 Marcovich 

Of the rituals associated with the Phrygian cults, two, the taurobolium 
and the self-gelding of the galli, have for obvious reasons attracted most 
scholarly attention, just as they attracted various types of obloquy in 
antiquity' 39 In my view the issue of gelding has too often been treated 
in rather anecdotal, formalist or phenomenological terms. In social 
anthropology, irreversible physical mutilation usually marks different 
stages in the process of social maturation (Leach 1976, 61). When such 
practices are taken over by a different culture, their precise connota- 



1 1 1 BC (Becher 1991, 163). Bayet 1957, 152 n. 14 noted this discovery, without under- 
standing its significance. Lambrechts, who had earlier (Lambrechts 1952b) committed 
himself to the claim that Attis was not a god at Rome before the supposed reforms 
of Antoninus Pius, steadfastly refused to admit that the new archaeological evidence 
proved anything about his status in this early period (1962; 1967). However, as Roller 
1999, 277 observes: "[These terracottas] demonstrate that Attis was an essential part 
of the Mother's cult from its inception in Rome, far more prominent there than in the 
eastern Mediterranean region"; see also Lancellotti 2002, 77—9. Attis may not have 
been part of the state cult at this time, but he was evidently considered a god especially 
protective of children by the population. 

158 The popularity of the story of Claudia Quinta might be read as a continuing 
mythical legitimation of this family orientation, cf Bomer 1964, 146f 

159 Sanders 1972 offers an excellent collection of material on the galli; cf. on the 
Hellenistic evidence, Giammarco Razzano 1982. 



THE RITUAL SYSTEMS 247 

tions tend to get lost. In this case, it was only in the fourth century 
AD that intellectuals were in a position to develop schemes in terms 
of which such practices could re-acquire their symbolic value. It was 
only then that the integrative function of rites of violence could be 
understood. The case of Isis was different, since persecution in that 
case was motivated by conflicts within the dominant class at Rome, 
not by the strangeness of the Egyptian rites. 140 

Philippe Borgeaud has recently argued that the galli practised (or 
suffered) total ablation, that is, the removal of testes, scrotum and 
penis. 141 This seems to me highly unlikely. In the first place, historically, 
total ablation is extremely uncommon, since the loss of blood involved 
meant that it was almost invariably fatal. It was practised in Qing 
China because radical sexlessness was required of court eunuchs, and 
there special surgical techniques were developed in order to increase 
the chances of survival; brutal ablation was practised on the so-called 
Black Eunuchs brought from the Horn of Africa to Islamic courts, but 
the death-rate was extremely high. 142 If, like Borgeaud (1996, 201 n. 73 
= 2004, 43), we suppose the usual model, of the man ablating himself 
in a state of frenzied possession, death by loss of blood would have been 
the inevitable and invariable consequence. It seems improbable that 



110 Cf. Takacs 1995, though in my view she fails to pay sufficient attention to conflicts 
between the classes and within the elite. 

111 Borgeaud 1996, 201 n. 74 (= 2004, 148 n. 74), citing G. Vorberg, Glossarium eroticum 
(Stuttgart 1932) 169. There are a number of relevant ancient words, e.g. ekteuveiv, 
zvvovxoq, cnrdStflv, aaGpoi, ydXXot, OXtfiiai;, OXaSiai;, castrare, castratus, spado, gallus, bagoas, 
megabyz(x)us, thlibias, thlasias. If the meaning of some is fairly exact (GXifSiai;, 0Xa8iai; 
mean resp. a man whose testicles have been rendered useless by being ligatured, by 
being crushed with a stone, cf. Plutarch, Micias 13) the others are quite imprecise. As 
Ulpian observed: spadonum generalis appellatio est: quo nomine tarn hi, qui natura spadones sunt 
(i.e. 'natural eunuchs', with hypogonadism due to chromosomal abnormalities such as 
Klinefelter's syndrome), item thibliae thlasiae, sed et si quod alius genus spadonum est, continetur 
(ap. Dig. 50.16.128). Elsewhere, in relation to dowry-rights, he opines that if a spado 
marries distinguendum arbitror castratus fuerit, necne (Dig. 23.3.39.1). Unlike other slaves, 
castrated men could neither 'marry' nor enjoy quasi-dowries; but any other kind of 
spado of this status might do so. In Roman Egypt, eunuchs evidently might make a 
testament, but their right to dispose of their property was tightly restricted: Gnomon of 
the Idios Logos §112 (BGU 5 p. 39). 

112 Cf. K.M. Ringrose, Eunuchs in Historical Perspective, History Compass 5.2 (2007) 
495—506. Celsus remarks on the number of blood-vessels in this region: Med. 7.18.2; 
19.4; and on the risk of heavy blood loss if the larger vessels leading to the scrotum 
are severed during an operation on the testes: ibid. 7.19.3: ne (venae) periculose sanguinem 
fundant. The real danger however lies in severing the large blood-vessels supplying the 
penis. A surgeon of course knew how to ligature or cauterise (small) blood-vessels: 
8.19.8; 22.1; 22.5. 



248 CHAPTER FOUR 

individuals would have volunteered to run such a risk, whatever their 
state of mind. Secondly, the loss of the penis means that the urethra 
has to be kept open by means of a tube to prevent it from occluding; 
and this of course would have led under ancient sanitary conditions 
to constant urinary infection, and worse. If this had been the case in 
the Graeco-Roman world, we would surely have heard about it in the 
medical writers. 143 Pliny twice uses an odd phrase, citra perniciem, when 
talking about the Galli: in one of these, he says Matris deum sacerdotes, 
qui Galli vocantur, virilitatem amputare nee aliter (i.e. with a sherd of Samian 
ware) citra perniciem. , 144 Pernicies, a common word in Celsus, here clearly 
means 'danger', even 'life-threatening danger' rather than simply 'dam- 
age'. 143 This implies that the issue of safety, and the risks involved in 
self-castration, were indeed thematised among these men. One could not 
have used such an expression if total ablation had been involved, since, 
Samian ware here or there, it was always extremely dangerous. 146 

If we look more closely at the language used about the Galli in this 
connection, it dissolves. Take for example Juvenal's well-known phrase: 
ingens semivir. . . mollia qui rapta secuit genitalia testa iam pridem (6.513-15). 14 ' 
Such an expression is completely non-specific: we can see from the 
usage of Pliny and other Latin writers, which of course often applies 
to animals, that genitalia could be used freely to mean male genitals or 
female ones (in that case either internal or external); when used of the 
Jews, the word clearly denotes the penis; of the beaver or the lion, the 



143 Celsus, for example, discusses urethral catheters, made of bronze, at Med. 7.26.1, 
but does not mention the possibility of an ablated male patient. 

111 Pliny, HM 35.165; cf. 11.262: citra perniciem amputantibus Matris deum Gallis. In the 
passage quoted in the text, he cites a certain M. Caelius (unknown to RE; the Vitellius 
named there may be the profligate Q. Vitellius mentioned in Tac, Ann. 2.48.3 = RE 
Suppl. IX, 1962, 174 If. Vitellius no. 7g) for the claim that a sherd of Samian ware 
alone could prevent pernicies during self-castration. 

145 E.g. Celsus 1.3.6; 2.8.36; 43; 4.13.1 against 7.20.2. 

'"' The mythical account of Attis' death of course implies that he bled to death as 
a result of gelding himself. The sole item of iconographic evidence relevant to this 
question is a relief from the Attideum at Ostia that shows the knife and the severed 
scrotum lying between Attis' legs; presumably the original colouring showed the blood 
pouring from the wound (CCCA 3: 119 no. 384, pi. CCXXXIX; Rieger 2004, 284 
no. MMA 9 with p. 132 fig. 97b). The scrotum is just that: there is no penis attached. 
One could hardly wish for better evidence against Borgeaud's claim. 

147 Mollis of course hints both at the impotence associated with eunuchism and at 
pathic sexuality, cf. Martial 3.73.4f: mollem credere te virum volebam, sed rumor negat esse 
te cinaedum. The comic juxtaposition of genitalia testa draws on the obvious pun testis- 
testa, which in my view is the origin of the trope about the Samian sherd. This would 
covertly acknowledge that the Galli lacked only their scrota/testicles. 



THE RITUAL SYSTEMS 249 

scrotum/testes. 148 The real point is that eunuchism always excites com- 
ment. In antiquity, in relation to humans, it almost invariably evoked 
accusations of pathic sexuality: cinaedus is virtually the first idea evoked 
by the word gallus. 1 * 9 It is this automatic association that inspires the 
one text that seems clearly to state that the galli were totally ablated: 
Quid cum femineo tibi, Baetice galle, barathro? haec debet medios lambere lingua 
viros. abscisa est quare Samia tibi mentula testa, si tibi tam gratus, Baetice, cunnus 
erat? Castrandum caput est. . . (Martial 3.81.1— 5). 150 

In my view, Martial had no interest in our question, total ablation 
or not? He is playing with a series of conventional oppositions, inver- 
sions and equivalences: male — female, no longer male = quasi-female, 
vulva — anus, mouth — anus, tongue = penis, head = penis. The basic 
joke is that galli are acknowledged to be men, but their absence of 
genitals, and their wearing of clothes classified as 'female', defines them 
as non-men; non-men ought to be pathic homosexuals, but Baeticus is 
behaving like a proper man with a woman — albeit with his tongue. The 
only way to put things right is to chop his head off too. For this sequence 
to work, Baeticus must be deprived of his penis, not his testicles. The 
same assumption is needed in the joke against Natta, who practises 
fellatio on his muscular draucus, collatus cuigallus est Priapus (1 1.72). 151 In 



118 Cf. //j\"20.89; 22.140; 23.65, etc. (testes distinguished from genitalia, it is never clear 
whether this opposes male to female patients or testicles from penis); 23.74, 24.1 17, 188 
etc. (simply genitalibus); 7.69 (female external); 1 1.262 (enlarged clitoris in some women); 
Jews: Tacitus, Hist. 5.5.2; HA Hadr. 14.2; = camel's penis, good for bowstrings: HN 
11.261; beaver: Apuleius, Met. 1.9; lions etc.: Pliny, HN 10.17.3; ass: Apul., Met. 7.23. 
Castrare applied to male animals, i.e. removal of testes: Pliny, HN 8.108; 198; 11.168; 
24.72; 30.41; 148 (of humans: 9.80; 11.230; 269); to female animals, by excision of 
the vulva (i.e. uterus): 8.68 (war-camels); 209 (sows and camels). 

1,9 The association gives point to the entire scenario of Martial 3.91, where a band 
of galli is imagined as trying to turn the boy-friend of a veteran into one of themselves 
by castrating him; and to the fragmentary 'Narrative about Iolaus' published as POxy 
3010 by Peter Parsons, dated first half of II 1 '. The unnamed speaker has become a 
gallus in order to help Iolaus pretend to be one as well so as to gain access to a woman; 
for some reason he addresses Iolaus and a Kivai8o<; together. 

150 "What concern have you, eunuch Baeticus, with the feminine abyss? This tongue 
of yours should be licking male middles. Why was your cock cut off with a Samian 
sherd if you were so fond of a cunt, Baeticus? Your head should be castrated..." 
(tr. D.R. Shackleton-Bailey). 

151 Shackleton-Bailey rightly however translates "compared to whom Priapus is a 
eunuch", since too much specificity would spoil the joke. It is not meant to be pressed 
to produce the meaning 'has no penis'; the same applies to the phrase gallo. . . Priapo in 
1.35.15; cf. at nunc pro cervo (i.e. a runaway slave) mentula supposita est (3.91.12). None of 
the passages cited by Kay 1985, 225 on 1 1.72.2 in fact supports the conclusion about 
ablation (Horace, Sat. 2.45 is not even about Galli in any sense). In 9.2. 1 31. Cybele 



250 CHAPTER FOUR 

another epigram, 13.63, however, the joke depends on the gallus having 
a penis but no testicles. 152 In yet another, a spado is supposed to have a 
penis but be unable to use it (1 1.81). 1M Elsewhere, the gallus clearly just 
lacks his testicles (3.24.14: caper). There is no consistency here because 
such jokes are not concerned with the technicalities of castration, they 
are about the social meaning of unmanliness. 154 The social knowledge 
was that the galli were eunuchs and dressed in 'women's' clothes; in 
ordinary daily experience, as the new curse-tablets from the joint temple 
of Isis and the Mater Magna in Mainz show, it was the self-castigation 
and self-laceration by the galli and bellonarii in the streets, and the spat- 
tering of blood, that people noticed, even though they were aware of 
the fact of castration. 135 In view of these considerations, I conclude 
that the galli normally simply lacked testicles and scrotum. Although 
the ideal was to perform this operation in a frenzy, on the model of 



herself is supposed to produce cinaedi; which allows Martial to slide to his point, that 
their bad-mannered host should have been paid back by having his penis cut off 

152 Ne nimis exhausto macresceret inguine gallus (i.e. a rooster), amisit testes. Nunc mihi gallus 
(i.e. a eunuch) erit; cf. 13.64. Likewise in 2.45, the fact that Glyptus can now have had 
himself circumcised (praecisa mentula) shows that, if he were already a gallus, he must 
still have possessed a penis. But it is obvious that in many of these scoptic poems gallus 
has lost its technical meaning of a mendicant follower of Cybele; the same process is 
clear in late Hellenistic Greek: Epist. 'Diogen. Sinop.' 11 (p. 16 Miiseler): yaXXoiq Kai 
Kivai8oX6yoi<; |j,£xa8i8a>Gi (date uncertain, between late IP and II P , but the technical 
sense appears in the next sentence); Josephus, AJ 4 §290 (list of 'Mosaic' laws): y6.XXox>q 
eKTpHieaGcu . . . 8fjXov y«P, w? tfl? W^xA? auxoii; xeGriXviauevrii; |jexeKoo|xriaav'co 7tp6<; 
xoCxo Kai to aa>|ia (this is the sole occurrence of the word in J.); in the Gnomon of the 
Idios Logos §112 (BGU 5 p. 39, c. 150 CE), yaXXoi are paired with craGpoi, impotent 
men, without any reference to the technical meaning. 

153 Cf. Kay 1985, 239 ad loc. on the popularity of eunuchs as sexual partners, e.g. 
Caelia in 6.67, who only has eunuchos as slaves, because vultfutui nee parere. 

1,1 Cf. Appendix proverbiorum 1 §67 (CPG 1 p. 389) s.v. yaXXwzl X£|j.£iv yaXXoi 
KaXoCvrai f|...Ti oil neTcmeTtTcbKaaiv eiq exepav ipuoiv; cf. Macarius 2 §92 (CPG 2 
p. 152f); 6 aXXoq xi<; ytvoiisvoq kou uexaPaXtbv xr\v (pioatv zh\ avSpcov, Kai ouxe avip 
aiv oike yuvfi: Etym. mag. p. 220.22-24 Gaisford; 6 aXXoc, xiq ye.v6iie.voc, Kai uexa|3aXa)V 
xr|v cp-uatv it, avSpcov: Etym. Gudianum, p. 296. 6f de Stefani. 

155 Cf. Blansdorf 2005, 674 no. 182,4 = 2008, Appdx., no. 16 11. 9-11: quomodo 
galli Bellonari magalfi] sibi sanguinem ferventem fundunt, frigidfum] ad terram venit...; 17 A6: 
quomodo galli se secarunt. . . . Earlier in no. 16, however, though the self-laceration comes 
first, reference is also made to self-castration: 16 11. 5f: quomodo galli se secant et praecidunt 
virilia sua. . . . Only in 18 11. 3: ita ut galli Bellonarive absciderunt concideruntoe se... does the 
self-castration come first. This is logical, since in order to become self-lacerating Galli 
they had to have castrated themselves/been castrated in the past. In these latter cases, 
the reference to the self-castration does not really fit the context as well as spattered 
blood does: it is extraneous social knowledge forcing its way into the curse through 
the metonymic chain Mater Magna — gallus — de-manned — cinaedus — weakness — death. 
Epist. 'Diogen. Sinop.' 11 (p. 16 Miiseler) also notes the attraction of Galli as a street- 
spectacle. 



THE RITUAL SYSTEMS 251 

Attis in the myth, 156 we must certainly allow for a variety of practices, 
including consecration of boys during childhood in the region of Pes- 
sinus (and probably therefore ligaturing the testes), and perhaps even 
surgical removal. 157 

We may now move on to the name. We do not know how or where 
the term Gallus originated. Sayce argued plausibly that the word 
derives from the Hittite iskallis. 1 ™ The earliest classical references to 
the institution are by Arcesilaus of Pitane, head of the Academy in 
the early third-century BC (cited in the epigraph to this section), 159 and 
Callimachus, whose fragment is the earliest known effort to write in gal- 
liambics (the metre of Catullus 63 and Maecenas' imitation), believed 
to be an imitation or caique on the rhythms of Phrygian dance: yakXax 
ur|xp6<; 6pzit\c, (ptA,69i)paoi SpouaSec;, odq evxea naxajeixai kou %dA,K£oc 
Kp6xocA,cx. 160 Both early references thus highlight the motif of sexual 



156 Cf. the paroemiographic entries s.v. yaXXtcrci T£|j.eTv cited in n. 130. LSJ 9 s.v. 
ya/VXtcm makes an effort to be jolly here, and translates 'cut the Gordian knot'; but 
the expression literally means 'to slice like a Gallus', i.e. without much reckoning of 
the consequences. 

157 To judge from Med. 7.19, a practitioner such as Celsus would have been perfectly 
capable of performing such an operation; Hadrian's rescript assumes that slaves will 
have been castrated by a doctor (see below). In such cases, total ablation will have 
been a serious possibility since the risk of the patient bleeding to death would have 
been much smaller. From the time of Domitian (Suetonius, Dom. 7; Dig. 48.8.6, PAD 
83), however, there were several imperial rescripts forbidding the practice; Hadrian 
indeed treated it as a form of murder (Dig. 48.8.4.2), cf. A. Watson, Roman Slave Law 
(Baltimore and London 1987) 122f No doubt however the castration of the galli fell, 
as in the case of the Jews, under the rubric cuius regit), eius superstitio, formulated, though 
of course not in those words, by Antoninus Pius (ap. Dig. 12.2.5.1). 

158 HA. Sayce, Kybele and GaUos in the Hittite Texts, CR 42 (1928) 161-63, followed 
by Walde-Hofmann s.v. gallus, 2 (1 p. 581). Greek etymological dictionaries assume 
a foreign origin: Boisacq and Frisk omit the word entirely, clearly because for them it 
was a loan-word; cf. Chantraine s.v. (1 p. 208). 

159 Note however the reservations expressed by Borgeaud 1996, 215 n. 43 = 2004, 
161 n. 43. I see no reason myself to doubt the tradition; given his theory about the 
Gauls, Borgeaud has every reason to doubt the ascription of the joke to Arcesilaus. 

160 'Fleeting-running Gallai of the mountain-Mother, lovers of the thyrsus, whose 
instruments and bronze finger-clappers fill the air with their noise', frg. 761 inc. auct. 
Pfeiffer = frag, adesp. Alex. 9 (2.6, p. 213 DiehP). The ascription of the lines to Callima- 
chus has rightly been defended by A. Dale, Galliambics by Callimachus, CQ,57 (2007) 
775—81 against the arguments of D. Mulroy, Hephaestion and Catullus 63, Phoenix 30 
(1976) 61-72 and E. Courtney, Three Poems by Catullus, BICS 32 (1985) 90f. As Pfeiffer 
notes, Catullus picks up the deliberate shift of gender of ydXJuxt in 63.12 etc. The 
apparent gender of (piX66l)pooi heightens the ambiguity. Pfeiffer's commentary explores 
the complex intertextuality of the passage, cf. Euripides, Helen 130 If, where the crasis 
of Cybele-Demeter-Dionysus already occurs, with Kannicht's excellent commentary 
(2: 327-59). Baslez 2004, 242f argues that the passage alludes to the fact that in Phrygia 
the ecstatic whirling dance of the galli was also practised by women. 



252 CHAPTER FOUR 

ambiguity. The usual ancient explanation was that the name derived 
from the river on whose banks Attis emasculated himself: 161 according 
to Alexander Polyhistor, in the first century BC, its name was Terias, 
but Attis (or rather Gallos) called it rdAAoc;. 162 The vulgate, deriving 
from Callimachus, claimed that the waters of this river had the property 
of making one possessed. 163 Quite which river in this part of Anatolia 
was intended is an insoluble problem. 164 More recently however efforts 
have been made to revive Jerome's etymology which connected it with 
the name of the Gauls, in Greek Takaxai, who invaded Asia Minor 
under Leonnorius and Lutarius in 278 BC and spent years plundering 
Seleucid territory even after their defeat at the 'Battle of the Elephants' 
in c. 268. 165 Both E.N. Lane and Philippe Borgeaud have argued in 
favour of linking the Gauls' invasion, the name of the river Gallus and 
the sacerdotal term. 166 Borgeaud in particular has stressed the analogies 
between the rough priests of the Great Mother and the comportment 
of the Gauls in battle (Livy 38.17.4), which will not have been lost on 



161 Cf. Appendix proverbiorum 1 §67 (CPG 1 p. 389) s.v. yaXXiaxl xeneiv: ydXXoi 
KaXoCvtat oi cntoxex|j,ri|j.evoi, f\ anb TaXkov xou 7tora|j,0'6 . . .; Macarius 2 §92 (CPG 2 
p. 152fi). This etymology was accepted straightforwardly by Ernout-Meillet s.v. gallus 
(2) in ed. 1 (1932) 392, but in ed. 2 it is distanced: "Les Latins le derivent...' p. 267. 

162 Alexander Polyhistor, FGrH 273 F74 (from the Description of Phrygia, Bk 3): 
Koci oxi xov YaXkov kou xov 'Axtiv ajtoKoyou xa ouSoia, k<x! xov |a,£v YakXov iXQexv inl 
xov Triptav itoxauov kou otKrjaou koi xov 7ioxa|j6v TaXXov KaAioar an' ekeivoi) yap 
xotx; xeuvouevoix; xa aiSoia ydXXotx; KaXoGatv. On the problem of this other person 
named Gallos, whom Julian, Or. 5.165b evidently identified or muddled up with Attis, 
see Jacoby ad loc. (3A Komm. p. 285), citing the (fictitious) Cappadocian king-list ap. 
Diodorus Sic. 31.19 = Photius, 382a30 Bekker = vol. 6 p. 141 Henry, where Gallus is 
the son of Queen Atossa and Pharnaces of Cappadocia. 

" ,:! Callimachus, Hepi 0oru|j,ao"icov (frg. 411 Pfeifler), from Pliny, HN 31.9; cf. Ovid, 
Fasti 4.363f. 

164 It is simplest to consult the RE entries s.v. Gallos nos. 1—3, with Barrington, maps 
62, F5 and G3; and 86, A3; also Cumont 1912, 674f. 

165 Jerome, In Osee 1.4.14, p. 44 Adriaen (CCSL 76): Hi sunt quos hodie Romae 

matri. . . servientes, Gallos, vocant, eo quod de hac gente Romani truncatos libidine sacerdotes illius 

manciparint. Propterea autem Gallorum gentis homines effeminantur, ut qui urbem Romani ceperant, 
hac feriantur ignominia. On the date (traditionally 277 BC), see M. Worrle, Antiochos I., 
Achaios der Altere und die Galater: Eine neue Inschrift in Denizli, Chiron 5 (1975) 
59-87; E. Will, Histoire politique du monde hellenistique, I (Nancy 1979) 143f; Mitchell 
1993, 18. Mommsen 1889, 1: 869: "und wenn auch die Regierung noch (in 205 BC) 
streng darauf hielt, daB die Castratenpriester der neuen Gotter Kelten (Galli), wie sie 
hieBen, auch blieben". 

166 E.N. Lane, The Name of Cybele's priests, the Galloi, in idem 1996, 117-33; 
Borgeaud 1996, 119f. (= 2004, 80-2). 



THE RITUAL SYSTEMS 253 

Roman readers. lr,/ However Cumont was surely right to point out that 
the word is only found in Greek after the establishment of the Seleucid 
Empire had made access to the temple-state of Pessinus practicable, 
despite the facts that a) the cult of the Great Mother (or forms of 
it) was known in Greece considerably earlier, and b) as was already 
assumed in antiquity, it must be calqued upon a Phrygian word. 168 It 
seems impossible to connect Greek TaXdiai with ydAAxn; it is only in 
Latin that there could be a play on the words Galltis, Gaul, and gallus, 
rooster, and so a nudge towards the notion of 'capon' {caper)} 69 

Although Borgeaud concludes from this that the Greek word yaXkoc, 
must be a borrowing from the Latin gallus after the arrival of the Gala- 
tians in Asia Minor (1996, 121 = 2004, 81), the more natural conclusion 
is that self-emasculation was a Phrygian institution prior to the arrival 
of the Tolistobogii in the region of Pessinus around 266. 17 ° The temple 



167 Although he rejects the idea that the Mater Magna might have been considered 
Gaulish (p. 120f. = 2004, 8 If.), he does wonder whether the Gauls might not have 
been disposed to worship Cybele in Anatolia because they had known the Mritrip Getov, 
an Archaic version of the same goddess, in the neighbourhood of Massilia (p. 121 — 
2004, 82). We may note that the image of the Gauls as tall, hairy, uncivilised brutes 
belongs to literary stereotype and has nothing to do with reality, cf. Darbyshire et al. 
2000; S. Mitchell ap. Erskine 2003, 280-93. 

168 Cumont 1912, 675. Casaubon was already of this opinion: "asiatica enim vox est, 
quae tov OOTOKXmov significat, neque a Galliae populis ea notio manavit" (ed. Lampridius, 
Heliog. 7, p. 806; cited by Lobeck 1829, 659 n.a [on p. 660]). Hesychius defines the 
probably related word ya.XXa.poc, as •J'pvyiaKov 6vo|j.a (1 p. 361 Latte); it occurs in two 
Greek inscriptions in the sense 'Dionysiac initiate': G. Dunst, TaXXapm, Zeitschrift fur 
Vergleichende Sprachforschung 78 (1963) 147-53; cf. LSJ 9 Suppl. p. 33. There exists a slight 
possibility, however, that it is after all Greek: there was an Aeolic form of the word 
i\Xoq, nail-head, spelled with a gamma, ydXXot (pi.) (Hesychius s.v.); the word had a 
second meaning, 'wart', 'callus'. The word ydXA-oi; might therefore have originated 
as a derogatory reference in that area of Greek settlement to the castration-scar, and 
thence, by synecdoche, applied to the men who did this to themselves. 

11,9 Walde-Hofmann s.v. gallus, 1 (1 p. 580f.) assume that the Latin word is a bor- 
rowing "aus einer vorasiatischen Sprache", citing Gk mXXoxov, cock's comb; KaXXa'iq, 
hen, and therefore unrelated to the word for a Gaul. Ernout-Meillet s.v. gallus, 1 
(p. 266) do toy — unconvincingly — with that idea (citing Gk. |^f|5oi;, 7t£patKov) but 
prefer onomatopoeia. The best-known example is the modius of the archigallus of 
Ostia M. Modius Maxximus (sic), now in the Vatican Museums: CIL XIV 385 — ILS 
4162 = CCCA 3: 123f. no. 395 pi. CCXLV - AE 1998: 275a,b = Rieger 2004, 146f. 
no. MM2 with figs. 119a,b; cf. Vermaseren 1977 pi. 64 (late IP-early IIP). We may 
perhaps take it that the Romans took a certain satisfaction in turning their former 
conquerors into capons. For Maxximus, however, there can have been no such joke: 
in his case the rooster is a mere rebus. 

170 Cf. Giammarco Razzano 1982. On the Tolistobogii, cf. Mitchell 1993, 55-58; 
K. Strobel, Die Galater: Geschichte und Eigenart der keltischen Staatenbildung auf dem Boden des 
hellenistischen Kleinasien, 1 (Berlin 1996) 252-57; Darbyshire et al. 2000, 79. The Greeks 



254 CHAPTER FOUR 

after all was reputed to have built by King Midas (Diod. Sic. 3.59.8). 
Two opposing accounts have been offered. One, exemplified by Frazer, 
is that the practice was an established ritual requirement which was 
then legitimated by the story of the self-castration of Attis. 1 ' 1 There is 
also an Euhemerist version, which assumes the divinisation of an actual 
human-being who had indeed castrated himself." 2 The alternative is 
that the fantasy preceded the institution: once the myth exists, people 
will start living it out. The vulgate assumes a model according to which 
the followers of Attis fell into a frenzy by drinking the waters of the 
river Gallus, just as Attis is said to have been frenzied (uaveiq) when he 
castrated himself. 1 ' 3 The theme is picked up by the Apologists, in the 
context of their hostile and highly implausible theories of the meaning 
of self-castration. 174 

None of these options seems very attractive. Thanks to recent work, 
we are now in a better position to understand the significance and per- 
sistence of the self-gelding of the Galli in the cultural context of the 



had of course long been familiar with the emasculate Megabyzus, the high-priest of 
Artemis of Ephesus. 

1/1 Frazer 1914, 1: 265: "The story of the self-mutilation of Attis is clearly an 
attempt to account for the self-mutilation of his priests". The discussion of Loisy 
1930, 93 focuses even more on raising doubts about the origin and original purpose 
of the self-mutilation. 

172 This sort of approach is surprisingly common in the wilder regions of scholar- 
ship, particularly in the case of Adonis. For example, Gottfried von Liicken, Beitrage z.u 
Geschichte, Kultur und Religion des Alten Orients (Baden-Baden 1971) 177—79, claimed that 
the son of the 10th cent. King Elibaal of Byblos, named Adonis, was the historical 
basis of the legend of the lover of Cybele. Sergio Ribichini, Adonis. Aspetti 'orientals 
di un mito greco. Centro di Studio per la Civilta Fenicia e Punica (Rome 1981) righdy 
makes the point that, even if this were true, it would not help us to understand Adonis' 
religious and cultural significance. He does not however make the point about euhem- 
erism being alive and well. In the case of Attis, the historical candidate is the son of 
another King, Midas (Frazer 1914, 1: 286). Lancellotti 2002, 16-60 revives this theory 
in a different guise. 

173 River water: Ovid, Fasti 4.361—66 with Bomer 2: 223; Festus, De verb.sign. s.v. 
Galli, p. 84.26f. Lindsay: dicti sunt a flumine, cui nomen est Gallo; quia qui ex eo biberint, in 
hoc furere incipiant, ut se privent virilitatis parte; cf. Herodian 1.11.2 etc.; Attis in a frenzy: 
Pausanias 7.17.12; cf. Julian, Or. 5, 167d with Ugenti ad loc. (p. 81); Arnobius, Adv. 
nat. 5.7 (= Sanzi 2003, 281, Cybele no. 39 .3): furiarum et ipse iam plenus, perbacchatus 
iactatus proicit se tandem et sub pini arbore genitalia sibi desecat. There were also euhemerist 
versions in which Cybele herself goes mad: Diod. Sic. 3.59.1; Firmicus Maternus, De 
errore 3.1 (with Turcan ad loc, p. 190); Vat. Myth. 1.225 Dain (jealousy); 3.28 Zorzetti- 
Berlioz (hatred). 

174 Cf. Sanders, 1978; also 1972, 1015-24. 



THE RITUAL SYSTEMS 255 

Roman empire. 1/S In the previous chapter I emphasised some aspects 
which will have contributed to understanding the underlying assump- 
tions of the myth in this connection; but here we can enlarge the scope 
of the enquiry and attempt to locate it within the entire complex of the 
paradigm castrating-mother/boy-lover. 176 The galli formed a specialised 
group in the service of the Great Mother, a group of priests whose 
self-gelding was rewarded by their control over the cult. 17 ' Excision of 
their genitals was an indispensable condition for this privilege. It was 
also a sign of their exclusive devotion to the cult of the Mother. In other 
words, renunciation of their physical integrity involved renunciation of 



175 Note the good survey of previous positions, and their intellectual backgrounds, 
in Sodergard 1993. There has been a good deal of useful historical work on different 
aspects of the phenomenon, e.g. K. Hopkins, The Political Power of Eunuchs, in Con- 
querors and Slaves (Cambridge 1974) 1 7 2—9 6 (orig. 1963); P. Guyot, Eunuchen als Sklaven 
und Freigelassene in der griechische-rbmischen Antike. Stuttgarter Beitrage 14 (Stuttgart 1980); 
S. Nanda, Neither Man nor Woman: The hijra of India (New York 1990); S. Marmon, Eunuchs 
and Sacred Boundaries in Islamic Society (Oxford 1995); S.H. Thai, The Eunuchs in the Ming 
Dynasty (New York 1996); P. Scholz, Eunuchs and Castrati: A Cultural History (Princeton 
2001); M.S. Kuefler, The Manly Eunuch: Masculinity, Gender Ambiguity and Christian Ideology 
in Late Antiquity (Chicago 2001); K. Ringrose, The Perfect Servant: Eunuchs and the Social 
Construction of Gender in the Byzantine Empire (Chicago 2003). 

176 I am attracted by the idea of working Borgeaud's Freudian reading (2004, lOOf.) 
into an historical account. I also subscribe to his view that the castrated status of the 
galli was essential to the religious meaning of the institution. Marie-Francoise Baslez 
has however argued on the basis of the Hellenistic evidence that the Anatolian galli 
were not necessarily castrated: their primary role, parallel to shamanic and near-eastern 
(specifically Hittite) frenetic dancing, was to utter prophecies by dancing, or whirling, 
themselves into a trance. It is only the later (Roman) sources that harp on the motif 
of castration (Baslez 2004, 243-5). I find this quite unconvincing: it is not at all clear 
that the Hellenistic epigrams she cites (many of them much later than she claims) are 
any less 'literary', viz. based on selected elements of a complex ritual actuality, than 
the Roman "stereotypes caricaturaux" she deplores. Moreover, she completely ignores 
the wider motif of blood-letting, which is central to the specific symbolic role of the 
galli under the Principate. Her sub-title 'Images et realites' spuriously implies that we 
can get 'behind' the play of claims and representations to an historical 'truth'. 

177 I use the word 'priests' here loosely, since the galli held no official title. Thomas, 
1984 has argued that they were not 'priests'; Sodergard 1993, 174 prefers to call them 
'cultic personnel' in order to avoid the Christian connotations of 'priest'. It is true that 
Pliny uses the word sacerdotes at HM 35. 165 (see n. 144 above); but Schillinger 1979, 345 
n. 2, is surely right to distinguish between those cities, such as Herculaneum, where there 
was an official cult of the Mater Magna, and thus official civic sacerdotes, and the loose 
use of sacerdos, "dem pneumatischen PriesterbegrifF des Ostens entsprechend . . . ohne 
dabei einen offentlichen Status zu besitzen. Aus dem Titel sacerdos allein laBt sich 
nicht erkennen, mit welcher Art von Priestern wir es jeweils zu tun haben". He uses 
the term 'Berufskleriker' for the galli. I am not convinced by the view that understands 
the emasculation as a preparation for priesthood. 



256 CHAPTER FOUR 

their 'proper' place in the social order. 178 In Pessinus and elsewhere in 
Mysia-Phrygia-Galatia this renunciation was compensated for by the 
high social evaluation of the choice. However, in Greek and Roman 
eyes, mutilation of their own bodies made it impossible for them to be 
accepted as normal members of the social group: classical conceptions 
of normative maleness meant indeed that the gallus was understood as 
a sort of slave. 179 

That being so, the myth of Cybele and Attis, apart from giving rise 
to a specific organisational structure, emphasised a destructive model 
of parent-child relations that can surely be explained with reference to 
the historical formation in which the myth was created. Of course we 
have no direct insight into its origins, since we only know it through 
the prism of the Greek and Latin authors, who in turn elaborated it in 
their own socio-cultural terms. In this new context, real and imaginary, 
the model played a decisive role in the construction of the imaginaire 
of the incest-taboo, thereby influencing behaviour in the context of 
mother-son relationships, both at the level of fantasy and that of real 
action. The model of the devouring mother, perpetuated in the societies 
that received the cult of Cybele, acquired dramatic expression in ritual; 
and that ritual performance could well on occasion have provoked its 
perversion, namely physical enactment. 

It is thus not out of the question that the model inspired direct 
imitation on the part of individuals. As I have noted in relation to 
self-abasement in prayer, the cult certainly encouraged a sense of 
personal emotional dependence. If the goddess demanded castration 
as a condition for becoming a true servant, the imperative might be 
met by individuals who, given the appropriate economic, social and 
ideological conditions, were emotionally disposed to turn fantasy into 
reality. Ascetic mysticism, insistent music, hallucinogenic drugs, a suit- 
able instrument — the combination might easily trigger such an act. 180 



178 Cf. B.-M. Nasstrom, The Abhorrence of Love. Studies in Rituals and Mystic Aspects in 
Catullus' Poem of Attis (Uppsala 1989) makes good use of Victor Turner on this point. 
However I find the substitution of the castration by a symbolic hierogamy unconvinc- 
ing; I think that the Phrygian cults had both a hierogamy (as part of the initiation) and 
castration (marking one's permanent dedication to her service). See p. 280 below. 

1/9 Cf. Seneca, de superstitione (ap. Aug., de civ. Dei 6.10): tarn indecora honestis, tarn indigna 
libris, tarn dissimilia sanis, ut nemo fuerit dubitaturus furere eos, si cum paucioribus fuerunt; also 
Bomer 1957-63, 4: 26. 

180 Sodergard calls these the 'extrinsic' stimuli as opposed to the 'intrinsic', mainly 
the imitatio Attidis (1993, 188). In his view, such stimuli or motives produced effects 
both 'intrinsic' (such as the belief that one's sacrifice might make the earth fertile, 



THE RITUAL SYSTEMS 257 

The worshipper loses his virility, as demanded by the devouring Mother, 
but preserves his selfhood, which is neither lost nor taken over. The 
son reacts before it is too late, and so escapes with his life. The Mother 
does not want to kill him, indeed she begs Zeus to preserve his life, but 
demands his sex and gets it. She thus gains a faithful servant, whose 
celibacy is irrevocable — no backsliding, no betrayal — and in return she 
grants him her everlasting divine protection. The model of Attis has 
achieved its effect. 

At the same time, however, we can view the process historically. The 
cult of the Mater Magna was originally brought to Rome from outside; 
with it came a myth created in a quite different cultural milieu — hence 
the Romans' coolness towards it. Different interests were in play. The 
plebs' acceptance of the new cult offered a potential node of resis- 
tance to the aristocracy. For the latter, popular participation simply 
compounded the problem of the cult's outlandishness. On the other 
hand, Rome was not in a position to send Cybele back where she came 
from: she had entered the pomerium quite legally, had protected the 
city; Hannibal had been defeated. Since she, the goddess, had fulfilled 
her part of the bargain, the Romans could hardly renege on theirs. 
As a compromise, her right to prey upon young men was limited to 
a group of foreigners. The detested galli thus fulfilled a specific social 
role: thanks to them, the Romans escaped self-castration but enjoyed 
Cybele's protection. Rome may have scorned the galli, but she needed 
them; these new defenders of Rome were actually indispensable. 

The ritual system in its broad lines, and thus (self-)castration, survived 
within the cult at Rome. 181 Some later versions suggest how the myth 
was re-interpreted in this new context. Lucretius writes: 

gallos attribuunt, quia, numen qui violarint 
matris et ingrati genitoribus inventi sint 
significare volunt indignos esse putandos, 
vivam progeniem qui in oras luminis edant. 

de rerum natura 2.614— 17 182 



underwrite the growth-cycle, or guarantee ritual purity) and 'extrinsic' (such as mark- 
ing boundaries, maintaining the integrity of the group, or meeting the demands of 
the castrating Mother). 

181 So much so that gallus comes to mean simply 'eunuch' (e.g. Martial 1.35.16, gallo 
turpius est nihil Priapo), a sense not listed by OLD; cf. Richard 1966. 

182 "They give her (Cybele) eunuchs as attendant-priests, to signify that those who 
have offended the Mother's godhead and shown ingratitude to their parents must 



258 CHAPTER FOUR 

Here the self-gelding is read conventionally as a punishment for slighting 
the goddess and/or lack of filial piety. 183 By implication, the frenzy that 
led to the act is to be understood as sent directly by the deity in response 
to an offence against her. Clement of Alexandria cites a myth, albeit in 
connection with Zeus, who, in the shape of a bull, has raped Rhea/Deo, 
that seems to be a mythical elaboration of the same idea: 

The Phrygians celebrate these festivals in honour of Attis, Cybele and the 
Corybantes. They declare that Zeus tore off a ram's testicles, and threw 
them at Deo's breast (&q anoanaaaq 6 Zevq xov Kpiori xovq 8i8'6|ioi)<; 
(pepcov ev \ieaoiq eppiye xdiq koAjiok; xr\q Ar\ox>q), pretending that they 
were his own (a>q eocutov 8fj6ev eKte|i(bv), and that he was paying the 
penalty for raping her. 184 

It would be best not to press this passage too hard and read it, as is often 
done, as a direct mythical legitimation of a substitutive ritual involving 
the removal of a ram's testicles, elsewere known as the criobolium. 1 ^ It 
does however introduce the idea of (self-inflicted) punishment. The pas- 
sage thus goes beyond Lucretius in suggesting that one ancient reading 
of the significance of self-castration was to redeem oneself. 186 This in 
turn suggests a parallelism between self-castration and self-mutilation 
by means of the whip and double-bladed knives, also practised by the 
galli. [& ' Given the familiarity of whipping as a punishment of slaves, 



be counted unworthy to bring forth living children into the sunlit world..." (tr. R.E. 
Latham, adapted). 

183 Cf. Summers 1996, 355-57. 

184 Protrept. 2.15 = Sanzi 2003, 266f. Cybele no. 29.1; cf. Sfameni Gasparro 1985, 
77-82. 

183 So righdy Thomas 1984, 1523, against e.g. W. Schepelern, Der Montanismus und 
die phrygische Kulte. Habilschrift Copenhagen (Tubingen 1929) 116; Vermaseren 1977, 
105; on the context, see Borgeaud 1998, 190-4; 2004, 114-17. It seems obvious that 
the myth is linked to the widespread theme of self-castration to avoid false accusa- 
tion, the so-called Kombabos-theme (Lucian, De dea Syria 19—27 with Lightfoot 2003, 
384-402). 

186 Against Meslin 1985, 181, I doubt that voluntary self-castration was linked to 
hierogamy, i.e. sacred marriage (cf. Sfameni Gasparro 1985, 81). According to Clem- 
ent, the hierogamy took place after initiation (see p. 000 below), when the mystes was 
still capable of reproduction. Self-castration was notoriously linked to sexual inversion; 
as a sign of radical change and a return, with the deity, to primal unity, it confirmed 
the subject's status as servant. 

187 |iacmv tav JtoXrjaaTpdyaXov, i.e. with knuckle-bones plaited into the thongs: 
Erycius (?late I 1 '), Anth.Pal. 6.234 = Garland, Erucius no. x 1. 4 (on the correct spelling 
of the name, see Garland 2: 278); Plutarch, Adv. Coloten 33, 1 127c; cf. latus horreat flagello: 
Maecenas, frg. 5 Baehrens = 6 Lunderstedt = 5 Courtney; knives: Apuleius, Met. 8.27 
(composite fiction drawing on the galli and bellonarii). 



THE RITUAL SYSTEMS 259 

and of criminals before their execution, this can hardly have failed to 
be understood as a form of punishment, and is indeed explicitly taken 
as such by Apuleius. 188 

I have already pointed out that renunciation of their physical integ- 
rity by the galli involved renunciation of one's 'proper' place in the 
social order. The idea of corporal punishment expresses one aspect of 
this renunciation — of one's own free will subjecting oneself to slave- 
punishment. 189 Another is cross-dressing, by assuming the saffron stola, 
and wearing make-up, prominent ornaments, pendants, ear-rings, and 
finger-rings. This rejection of the masculine norm is linked directly, as 
we have already seen, to the repeated claim that by castrating themselves 
these adherents of Cybele (and other galli) have become male pseudo- 
women, i.e. catamites. 190 Lucian reports that the galli in the temple of 
Atargatis at Hierapolis wore female clothing and did women's work; 
they 'feminised' (0T|A/6vovtcxi), allegedly in imitation of Kombabos, who 
in this case was upset because a woman who was vainly besotted with 
him had killed herself (De dea Syria 27). 191 Such explicit effeminisation 
calls attention to the rejection of this-worldly gender-classification, and 
all that goes with it — marrying to continue the household, having a 
fixed abode, working for a living. Failure to work for a living and the 
rejection of a roof over one's head were the hall-marks of the notori- 
ous iitixpayupTai in Classical and Hellenistic Greece, who, constantly 
on the move, lived off what they could beg from locals. In Latin they 
were known as famuli; Cicero calls for them to be licenced for particular 



188 Met. 8.28: quasi contra fas sanctae religionis dissignasset aliquid, et insuper iustas poenas 
noxii facinoris ipse de se suis manibus exposcere (the speaker goes on to whip himself). On 
whipping as a punishment of slaves, cf. Plautus, Persa 282; Horace, Epod. 4.3; criminals 
(including condemned slaves): F. Hinard andJ.-C. Dumont (eds.), Libitina: Pompes funebres 
et supplices en Campanie a I'epoque d'Auguste (Paris 2003) 118. 

189 Seneca, de superstitione (ap. Aug., de civ. Dei 6.10) suggests that the blood was 
understood as a supplicatio, a propitiatory offering: se ipsi in templis contrucidant, vulneribus 
suis ac sanguine supplicant. 

i9o Vermaseren 1977, 97; on cross-dressing, see Beard 1994. On the uses and mean- 
ings of different colours in make-up, see J. Alvar and T. de la Vega, La ambiguedad 
cromatica en los misterios, in P. Ortega, M.J. Rodriguez y C.G. Wagner (eds.), Mujer, 
ideologia y poblacion. II Jornadas de roles sexualesy de genero, Serie ARYS 11 (Madrid 2000) 
49-60. 

191 0r|Xi)v6|i£VO(; inevitably also had a sexual connotation, e.g. Vettius Valens, Anthol. 
p. 7.26; 10.19; 12.1 Kroll. De dea Syria also mentions (§52) that the Hierapolitan galli 
were buried under tumuli of stones outside the walls — they have 'lost' their connection 
to their birth-families, and require a special form of 'unnatural' burial; some oddly 
unperceptive remarks in this connection by Lightfoot 2003, 509f. Baslez 2004, 236 
insists upon the fact that the Phrygian galli allowed their hair to grow long. 



260 CHAPTER FOUR 

days, no doubt in order to prevent their indiscriminate multiplication. 192 
To judge from literary sources, however, they were a familiar sight on 
the open road. 193 All of these differences assert membership in an ideal 
society where the rules are quite simply different from those of the 
dominant world; and explain the odd mixture of horror and fascination 
that the galli provoked in the minds of their contemporaries. We cannot 
even guess at the proportion of male Metroac worshippers who may 
have gelded themselves: slaves were not free to do so; the separation 
from the ordinary world of work and achievement was so radical that 
we can hardly imagine many freedmen taking such a step. Perhaps 
most were always indigenous Phrygians, born into a culture where the 
practice may have had a higher social valuation. 194 Yet the existence of 
even a few monuments that seem to represent such men, for example 
at Ostia, implies that some at least were relatively prosperous, and had 
not severed their ties with the normal world. For them the ideal society 
was ideational, not literal. 

It is this membership in an ideal society that is connoted by the 
occasional representation of the cista mystica in the context of the 
galli. One such basket is depicted apparently hanging from a nail on 
a (temple-) wall represented beside the full-dress gallus on a well-known 
relief from Lanuvium. 19S A similar cista is also once depicted at Attis' 
self-emasculation. 196 For what it is worth, the scholiast to Clement's 
account of the 'mysteries of Deo' claims that Attis' genitals were placed 



192 Cicero, Laws 2.22: praeter Idaeae matris famulos eosque iustis diebus ne quis stipem cogito; 
cf. Ovid, Fasti 4.350-2 Alton-Wormall-Courtney: 'Die', inquam 'parva cur stipe quaerat opes.' 
'Contulit aes populus, de quo delubra Metellus fecit' ait; Dandae mos stipis inde manet'. 

193 E.g. Martial 3.91.2: a veteran semiviro Cybeles cum grege iunxit iter, cf. Apuleius, 
Met. 8.26. 

194 And perhaps therefore castrated as children (see n. 157 above). Lancellotti 2002, 
98f. however urges that the operation was confined to adults rather than pre-adolescents. 

195 CCCA 3: 152f. no. 466, pis. CCXCVI-VII = Cumont 1929, pi. II. 1 = Helbig 4 
2: 25f. no. 1176 (Capitoline Museum inv. no. 1207, perhaps mid-II p ). It is uncertain 
whether the association between cista mystica and the Archigallus (that of M. Modius 
Maxximus [see n. 169 above] and on the sarcophagus found at Isola Sacra [CCCA 
3:140f. no. 446]) evokes their self-gelding, their link with the galli or their initiation. 

196 Cf. the hexagonal floor-stone from an unknown provenance in Gaul, now in the 
Bibliotheque Nationale, CCCA 5: 157f. no. 465, where the cista stands at the bottom 
of the scene, with Attis above (some doubt must however attach to the authenticity). 
It must however be said that the many altars that set out the symbols and instru- 
ments associated with the taurobolium, such as the series from Die {CCCA 5: 122f. 
nos. 359—64), do not include the cista. This perhaps indicates that the ritual was not 
understood as an initiation. 



THE RITUAL SYSTEMS 261 

in a cista. 191 This may be confirmed by the defixio from Mogontiacum 
(Mainz), mentioned in Chapter 2.1.b above, which associates Attis with 
Castor and Pollux and the cistae penetrales, "containers kept within (the 
temple)". 198 The expression occurs nowhere else as such in the context 
of Metroac cult, but was evidently perfectly familiar in Mainz. 199 If we 
stress the connection with Attis, the phrase might be used to support 
the idea that such containers mainly connoted the god's self-gelding and 
death, and hence the galli. Whatever their precise contents, however, 
they seem clearly to evoke the idea that self-gelding was 'mystery', a 
privileged means of gaining access to the goddess. 

c. The taurobolium/criobolium 

Along with self-castration, the taurobolium/criobolium is the most strik- 
ing ritual component of the Phrygian cult. 200 Traditionally, however, 
it has received a bad press. 201 This is largely due to the Christian poet 



197 Z Clement Alex. Protr. 2.19 — Hepding 1903, 32. This is however not confirmed 
by any epigraphic source. 

198 AE 2004: 1026, with Blansdorf 2004, 56; 2005, 680 no. 201 B36. 

199 Note also the sequence, unfortunately garbled, in another of these texts: quomodo 
galli se secarent. . . quomodo et sacrorum (?) deposierunt (i.e. the galli) in sancto, sic et tuam vitam (et) 
vaktudinem, Gemella: Blansdorf 2005, 678 no. 72, 3 = 2008 no. 17 11. 6-11. For per cistas 
penetrales, cf. Apuleius, Met. 6.2: perego te frugiferam tuam dexteram istam deprecor, per laetificas 
messium caerimonias, per tacita secreta cistarum. . . (Psyche to Demeter/Ceres). A problematic 
passage in HA Elagab. 7.2 claims that Antoninus danced ecstatically like the galli, and 
genitalia sibi devinxit et omnia fecit quae Galli facere solent, ablatumque sanctum in penetrale dei 
sui transtulit. Despite the grammatical mistake (genitalia... ablatum), the thought seems 
to be that it was his severed genitals that were placed in the shrine of Sol invictus 
Elagabalus, in the same way as the galli placed theirs in the shrine of Mater Magna. 
(D. Magie in the Loeb edition takes ablatum to be a reference to the statue of Cybele 
et alia sacra quae penitus habentur condita mentioned in the previous sentence, which seems 
rather forced.) The accusation must be nonsense (so too Schillinger 1979, 39 If.), since 
Varius Avitus was already castrated; but it implies that Marius Maximus knew where 
the genitals of the galli were in fact dedicated, or that it was common knowledge (on 
the Life, see Barnes 1970, 30f). 

2110 In what follows, I sometimes speak of the taurobolium alone; in these cases it 
should be understood that I am also referring, mutatis mutandis, to the criobolium. As far 
as is known, the only major difference between them was the cost of the animal and 
the number of people its meat could feed. In both cases the sacramental element was 
the ritual handling of the animal's severed scrotum, the vires, in an evident reference to 
Attis' self-castration. Borgeaud 1998 (which complements his account in 2004, 1 10—19) 
is a useful recent discussion of the issues. 

201 Cumont 1929, 67 is typical: "Le taurobole, qui cherche a satisfaire les aspirations 
les plus elevees de l'homme vers la purification spirituelle et l'immortalite, apparait 
comme une douche de sang qui fait songer a quelque orgie de cannibales". A brief 
survey of opinions in Marco Simon 1997, 310f. nn. 41f. 



262 CHAPTER FOUR 

Prudentius' account, at the very end of the fourth century AD, of a 
supposed ritual that takes place at Rome: a summus sacerdos stands in a 
hole beneath a timber-grid and is showered with the blood of a bull, 
whose thorax has been pierced by a spear, the venabulum. 202 This account 
is spoken — miraculously — by the Antiochene deacon-martyr Romanus 
after his tongue has been cut out by the executioner: blood is therefore 
literally all about. Romanus' claim was supposedly supported by pas- 
sages in the contemporary Christian vilification of pagan practice, the 
Carmen contra paganos and Firmicus Maternus. 203 In the late 1940s the 
Dutch scholar K.H.E. de Jong produced an imaginative drawing of 
Prudentius' shower-bath, reprinted by Vermaseren in 1977 as a more 
or less authentic reconstruction of the event. 204 Although everyone 
allowed that this 'evidence' was rather late, it was still accepted as 
reliable, as a historical fact, and fitted into an evolutionary scheme of the 
rite by the two most important discussions of the taurobolium of the late 
nineteen-sixties, by Jeremy Rutter and Robert Duthoy 205 What Cumont 
considered a primeval rite was thus transformed into the third stage 
of a long historical development and associated with the late-antique 
dedications, mainly by members of the so-called 'pagan reaction', in 
the sanctuary of the Phrygianum on the Vatican. 

All this however wonderfully illustrates the dangers of working from 
excerpts and unfamiliar sources. Recent work on the Peristephanon, 



202 Prudentius, Peristephanon 10.1011-50. As Rutter 1968, 240 allowed, Prudentius 
nowhere mentions the word 'taurobolium', but this has not dented confidence in the 
authenticity of the claim. 

203 Carmen contra paganos 57-62 (see the ed. and tr. by C. Markschies ap. R. Feldmeier 
and U. Heckel (eds.), Die Heiden: Juden, Christen und das Problem des Fremden [Tubingen 
1994] 325-77; also Martinez Maza 2000, 66-75). This brief confection of course is not 
a source 'independent' of Prudentius, even if it is not by him — it is actually transmitted 
as part of his corpus (cod. Par. Lat. 8084, fols. 156 r — 158 v ). Cf. also Firmicus Maternus, 
De errore 27.8— 28.1. J.M. Poinsotte, La presence de poemes anti-pa'iens anaonymes dans 
l'oeuvre de Prudence, REAug 28 (1982) 35-58 suggested that Prudentius was elaborat- 
ing on these lines of the Carmen. 

201 K.H.E. dejong, Oosters-hellenistische Mysterien (The Hague 1949) 86 = Vermaseren 
1977, 104 fig. 30 = Turcan 1992a, 56 fig. 1. See also the first edition of the present 
book, p. 199 fig. 47. Such images, albeit frankly reconstructions, have a reificatory 
power not dissimilar to that of 'documentary' photos or scientific drawings. 

205 Rutter 1968; Duthoy 1969; cf. Burkert 1979, 1 19 and 1987, 98 (a fact); Vermaseren 
1977, 101-7 (perhaps a fact); Turcan 1992a, 55-58; still quite uncritical, Sanzi 2006, 
168 n. 190. Acceptance of the historicity of the passage goes back to Graillot 1912, 
153-74, and indeed beyond. Duthoy's book received a devastating review by T.D. 
Barnes, Gnomon 43 (1971) 523f; even before it appeared, Duthoy himself had expressed 
some doubts about his sequence: Traces archeologiques de tauroboles a Zadar, Latomus 
27 (1968) 622-9 at 623-5; cf. Richard 1969, 666-8. 



THE RITUAL SYSTEMS 263 

and the Romanus section in particular, has suggested that its account 
of supposed pagan practice has no pretension whatever to historical 
accuracy or authenticity, and is intended merely as a foil to the marty- 
rological topoi. 206 The shedding of the martyr's blood, and his rhetorical 
defence of the Christian faith, are the poem's two central themes. 207 
Largely independent of this growing awareness among scholars of 
early Christianity of the Tendenz of Prudentius' poem, Neil McLynn 
has brilliantly exposed the weaknesses of the Rutter/Duthoy scheme, 
emphasizing that Prudentius otherwise shows very little knowledge 
of paganism (he only knows the stock themes, the apologetic cliches); 
that the Carmen and Firmicus Maternus do not support the shower of 
blood, or any similar detail; that the 'description' fits all too neatly with 
the rhetorical and thematic needs of the moment, after more than a 
thousand lines of rhodomontade; and that the fourth- century taurobo- 
lium was a private occasion that the altars and inscriptions turn into a 
highly public, celebratory event (McLynn 1996). 208 

McLynn's arguments can in fact be supported, at least in a nega- 
tive sense, by looking at the archaeological evidence. 209 An important 
prop to the Prudentian scenario — granted that it was the Prudentius 
passage which inspired its identification as a "fossa sanguinis" (the 
term is modern) in the first place — used to be a pit that Guido Calza 
discovered in the tower of the Sullan wall to the south of the temple 
of Bellona during the course of his re-excavation of the precinct of 
the Mater Magna at Ostia. Anna-Katharina Rieger has now shown 
conclusively that this 'pit', which was in a stuccoed room (hardly suit- 
able for a butcher's shop), was partly vaulted, and that the small, square 
entrance could be covered with (a) board(s). The 'pit' was deliberately 
built so as to extend almost to the water-table, which at this point is 



206 j p Petruccione, Prudentius' Use of Martyrological Topoi in Peristephanon (Ann Arbor 
1985); R. Henke, Der Romanushymnus des Peristephanon, JfAC 29 (1986) 59-65; 
A.-M. Palmer, Prudentius on the Martyrs (Oxford 1989); M. Roberts, Poetry and the Cult of 
the Martyrs: the Liber Peristephanon of Prudentius (Ann Arbor 1993). 

2117 W. Evenepoel, Le martyr dans le Liber peristephanon de Prudence, Sacris Erudiri: 
Jaarboek voor Godsdienstwetenschappen 36 (1996) 5-35. 

21,8 Borgeaud 1996, 167 = 2004, 117f. came to the same conclusion at the same 
time, albeit without the same range of argument; cf. Rieger 2004, 112. At the inter- 
national conference Les 'religions orientals' dans le monde gree et romain: cent ans apres Cumont 
(1906-2006), Rome, 16-18 November 2006, Mary Beard followed up McLynn's cri- 
tique with an analysis of the role played by the shedding of blood in Peristephanon 10 
(to appear in Bonnet, Pirenne-Delforge and Praet 2009). 

209 Rather oddly, McLynn himself fails to mention this issue. 



264 CHAPTER FOUR 

rather high, and must therefore have constantly filled with water. She 
concludes that it was simply a cistern used to fill the various basins 
that are scattered around the precinct, notably beside the steps of 
the temple of Mater Magna. The flight of steps into the cistern were 
intended simply to make it easier to collect water. 210 Inspired by Calza, 
H. von Petrikovits interpreted a cellar with two flights of steps found at 
Gepaplatz 3, Neuss-Gnadental in Nordrhein-Westf. (Roman Novaesium) 
as a "fossa sanguinis". There can be little doubt that this cellar was in, 
or near, some sort of sacred area in the SW corner of the vicus of the 
auxiliary fort. However the only temples found were those dedicated 
to Jupiter and the Matronae. 211 No reliable evidence of a connection 
to Mater Magna could be found, and Elmar Schwertheim has rightly 
commented: "Diese Anzeichen. . . sind. . .keineswegs zwingend" (1974, 
9 no. 5). The supposedly independent archaeological confirmation of 
the Prudentian scenario thus dissolves. 

The most important aspect of this demonstration, combined with 
earlier criticisms of the evolutionary scheme, is that we need no longer 
believe that there were any major shifts (other than local variation) in 
the Metroac version of the taurobolium/criobolium once its ritual pattern 
had been established in the mid-second century AD. We need simply 
accept that, once the ceremony entered the cult's ritual-system, it formed 
an organic part of it. 

On the other hand, the adoption of the taurobolium/criobolium as a 
Metroac ritual implies that similar modes were already available within 
the spectrum of rites known in the Empire, so that we do not neces- 
sarily have to recognise a taurobolic ritual behind every occurrence of 
the word or every image of a bull. This has long since been established 
for Anatolia, where various forms of non-Metroac xa\)po(36A,ia are 
known. 212 We should thus be wary of interpreting some altars listed 
in the relevant corpora as Metroac rather than as evidence for some 
restricted indigenous ritual: I am thinking, for example, of two reliefs 



210 Calza 1946, 186; 197£; Rieger 2004, 1 10-12 with figs. 82-83b. She also dates its 
construction to the late I p , which is far too early for the Prudentian scenario of the Rut- 
ter/Duthoy hypothesis. Calza 1946, 196f himself dated it II 1 ', to Rutter's discomfort. 

211 Photo and brief discussion in H.-G. Horn (ed.), Die Romer in Nordrhein-Westfalen 
(Stuttgart 1987) 588f 

212 "The taurobolion is a ritual preserving elements of bull-hunting": Burkert 1979, 119. 
Some of the Anatolian evidence regarding non-Metroac Kpio[36A,ia (ephebes catching 
rams and sacrificing them, e.g. OGIS 764 1. 27f.) and rarjpo|36Xt,a (e.g. IGR 4.494 etc., 
TAM 2.508, SEG 2: 727 = AE 2003: 1748 with L. Robert, Les gladiateurs dans I'Orient 
grec [Paris 1940] 316-18) is presented by Rutter 1968, 226-29. 



THE RITUAL SYSTEMS 265 

found at Sos del Rev Catolico, prov. Saragosa, which seem rather to 
be connected with an ancient pre-Roman cult. 213 

It is generally believed that the first Metroac taurobolia were performed 
under Antoninus Pius (AD 138-61). 214 The earliest surviving epigraphic 
evidence (though we do not have to assume, with Beaujeu, that it was in 
very fact also the first) 215 relates to a sacrifice undertaken in AD 160 at 
the goddess' behest for the elderly Antoninus Pius, his children and the 
colonia Claudia Lugdunensis in or at the vaticanum.™ The spot is clearly 
named from the site in Rome of the same name. 21 ' Specific reference 
is made to the new, or newly-reinforced, rules for the appointment of 
local municipal priests of Mater Magna. 218 Presumably the noble exta 



213 CCCA 5: 77f. nos. 21 If. ("possibly a taurobolium monument"); also those described 
by C. Aguarod and A. Mostalac, Nuevos hallazgos de aras taurobolicas en la provincia 
de Zaragoza, Homenaje Almagro, 3 (Madrid 1983) 311-329; cf. Alvar 1993c, 39f.; J.E 
Ubifia, Mater magna, Cybele and Attis in Roman Spain, in Lane 1996, 405—33 at 
417. I do not find Marco Simon's case for linking the Sos del Rey Catolico monument 
to Cybele convincing (Marco Simon 1997, 306—10). 

214 I accept Rutter's case (1968, 23 If.) for regarding CILX 1596 = /IS 4371 = Duthoy 
1969, 29 no. 50 = CCCA 4: 8 no. 11 (7 Oct. AD 134), from Puteoli, as essentially an 
eastern Mediterranean ceremony (it is directed to Venus Caelestis). 

2,3 The casual reference to the local vaticanum suggests a familiar ceremony. If it had 
been the first, we would have expected some explicit notice of the fact; Pompeia Phi- 
lumene, for example, states that she was the first woman [prima) to celebrate a taurobolium 
at Lactora/Lectoure (CIL XIII 504 = ILS 4121 = ILAq Lectoure 3). Schillinger 1979, 
368 assumes, probably righdy that private taurobolia had been performed earlier. 

216 CIL XIII 1751 = ILS 4131 = CCCA 5: 133f. n. 386, pi. CXXXI: Taurobolio 
Matris d.m. Id., quod factum est ex imperio Matris {d.} deum pro salute imperatoris Caes. T. Aeli 
Hadriani Antonini Aug. Pii p.p. liberorumque eius, et status coloniae Lugudun., L. Aemilius Carpus 
VIvir Aug., item dendrophorus, vires excepit, et a Vaticano transtulit, ara et bucranium suo inpendio 
consacravit, sacerdote Q Sammio Secundo ab XVviris occabo et corona exornato, cui sanctissimus 
ordo Lugdunens. perpetuitatem sacerdoti decrevit, App. Annio Atilio Bradua, 77 Clod. Vibio Vara cos. 
L.D.D.D. . . . There is a large bull's head with infulae in bas-relief on the stone which 
divides the section ending dendrophorus from that beginning vires excepit. Presumably 
the word bucranium in the text refers to this image; but it is not impossible that the 
bull's skull was also buried. On the right lateral face of the altar is a brief additional 
text: Cuius mesonyctium factum est V. id. Dec. (9 December). The nature of this nocturnal 
ceremony is obscure (cf. CIL VI 10098 = 33961 = ILS 5172); but the addition implies 
that the altar itself was erected well before the end of 160. 

2I/ Rutter 1968, 235 argued, as others before him had done, that the sacrifice must 
have taken place at Rome and the vires transported to Lugdunum. I accept Schillinger's 
argument (1979, 364f) against this interpretation, which is based on the attested exis- 
tence of a vaticanum in Moguntiacum/Mainz (CIL XIII 7281 = ILS 3805, AD 236); 
cf. also Fishwick 1967, 143—5. The earliest taurobolic altar from Rome is dated AD 
295 (CIL VI 505 = ILS 4143). 

218 This inscription is the earliest evidence for the regular method of appointing 
civic priests of Mater Magna: Q. Sammius Secundus, the supervising priest, had been 
first selected by the local municipal senate (ordo) and then confirmed by the XVviri 
sacris faciundis at Rome, who granted him the right to wear the insignia of such office, 



266 CHAPTER FOUR 

were, as usual in Roman custom, first boiled then burned on an altar 
for the gods, and the rest of the meat eaten in situ by the participants, 
distributed or sold to butchers. What was specific to the Metroac form of 
the taurobolium was the special treatment of the severed scrotum {vires). 219 
In view of the arrangements in the sacred area at Ostia, one of the 
best-known Metroac sites, where the only killing-altar was a perfectly 
normal one in front of her temple, we should probably conclude that 
vaticanum was simply a name for such an altar, taken to represent a 
notional reference-point at Rome. 220 

The sudden appearance of this type of sacrifice, at any rate epi- 
graphically speaking (all epigraphic evidence for the taurobolium after 160 
in the Latin-speaking Empire is related to the cult of the Mater Magna), 
seems to be associated with a readiness on the part of relatively wealthy 
municipal figures to combine public and private concerns. 221 One con- 
sideration is no doubt that, at least in some places, persons who under- 
took a taurobolium for the benefit of the state {pro salute imperatoris) after 
obtaining permission from the local archigallus (ex vaticinatione archigalli) 
could be excused duty as Mores, that is, be freed from the responsibility 
for looking after the affairs of minors (including the children of liberti) 
and certain categories of women, which might be quite onerous. 222 



a wreath and the occabus (a Greek word meaning 'arm-band' otherwise known only 
from late lexica: Hesychius s.v.; Etym. magn. p. 383.21 Gaisford). The fullest statement 
is CIL X 3698 = ILS 4175 (Cumae, AD 289). Schillinger rightly argued that this was 
intended to tighten up a previously laxer or more informal practice (1979, 358f). Note 
that the centre of the Roman cult, at any rate in this connection, has shifted from the 
Palatine temple to the Vaticanum; there is however one mid IIP 1 taurobolic altar that 
gives Mater Magna the epithets Phrygjiae Palatinae: AE 1910: 217 = 1924:26 = ILG.N 
518 (under Philippus Arabs). 

219 It is unknown whether the scrotum was severed while the animal was still alive, 
after it had been stunned, or after it had been bled to death. Those who believe that 
the sacrifice was substitutive will doubtless incline to the first possibility. 

220 The altar was built in the Antonine period — roughly contemporary with the first 
taurobolium inscriptions — and continued to be used into IIP 1 ; there was no comparable 
construction earlier; see CCCA 3: 109 under no. 362 with pi. CCXV; Rieger 2004, 94 
fig. 61; 97 fig. 62; 121 fig. 90c. (oddly enough, Rieger herself does not even mention 
it.) There is no sacrificial altar in front of any other temple or sacellum in the sacred 
precinct. It would thus be reasonable to infer that the taurobolia/criobolia, as well as 
other sacrifices, took place here. 

221 I cannot follow Borgeaud however in his estimate that "the taurobolium was an 
aristocratic ritual whose purpose was to guarantee the health and well-being of the 
imperial house and the city" (2004, 92). 

222 Fragmenta Vaticana §148, de excusatione: [Is] qui in Portu pro salute imperatoris sacrum facit 
ex vaticinatione archigalli, a tutelis excus[a]tur (= FIRA 2 2: 496). On the law relating to tutelae, 
see E. Sachers, s.v. Tutela (1,2), RE 7A2 (1948) 1497-1599. Rutter not only wrongly 



THE RITUAL SYSTEMS 267 

Although undated, this concession (which is actually limited to Portus at 
Ostia but may have been applied more widely — we frankly do not know) 
has generally been linked to a reform of the cult of the Mater Magna 
under Antoninus Pius, though scepticism about the existence of such 
a reform has recently been aired. 223 A taurobolium/criobolium was a rela- 
tively lavish undertaking: just to list the most obvious items, it involved 
producing a full-grown bull, or a ram, from one's estate or on the open 
market, defraying the costs of the sacrifice itself and the preparation 
of the accompanying feast, paying for the stone and the stone-mason; 
obtaining permission to erect the finished altar. 224 The sort of people, 
often curiales, who could afford such gestures were seriously interested 
in divine support for the social order that maintained their property 
and security, their isogamous marriages and their access to municipal 
honours. 223 The interests of wealthy freedmen were hardly different. 
For them it was important that the sacrifice took place in public. 226 
The very fact that we still possess at least eighty-five taurobolic altars 
from the 220-year period AD 160— c. 350 (excluding the remarkable 
twenty-two altars alone from Lactora/Lectoure in Aquitania) suggests 
that there were a good many in the western Empire prepared to use 



believed that the concession related to tax exemptions (mis-citing the text on p. 234) 
but breezily declared, "I think it is safe to assume that the benefits to the dedicant of 
a public taurobolium (i.e. remission of taxes) applied not only to the inhabitants of 
Portus, but to individuals throughout the Empire" (p. 236 n. 34). Nothing could be 
less certain. 

22! A reform of the cult under Antoninus Pius was first suggested by Lambrechts 
1952a, cf. M. van Doren, L'evolution des mysteres phrygiens a Rome, AntCl 22 (1953) 
79-88; it was Beaujeu 1955, 313-20 who linked the changes to the taurobolium. These 
claims are now treated as facts by most scholars, e.g. Rutter 1968, 235f; Schillinger 
1979, 352—73; Thomas 1984, 1522; R. Sierra, Taurobolio y culto imperial en la Galia 
Narbonense, in Alvar, Blanquez and Wagner 1995, 201-14; Martinez Maza 2000, 
68. Borgeaud however, perhaps rightly, prefers to emphasise the speculative nature of 
the inference by speaking of the "reforms" (e.g. 2004, 119). Rieger, while accepting 
that Mater Magna has a special place in Antonine coinage, argues that such 'reforms' 
were merely formal recognitions of gradual, 'spontaneous' changes in organisational 
and ritual practice (2004, 163-5, but see p. 172). I revert to this issue several times in 
the following pages. 

221 A cursive inscription on a brick from Sains-du-Nord (Belgica) has been read as 
the account for furnishing the meat required for a taurobolium (spelled torobol(ium)): 
AE 2001: 1398, cf. 1997: 1140). For what little it is worth, Carmen contra paganos 57f. 
contrasts the vast wealth of the celebrant (by implication, a pre-requisite in view of 
the expense of the taurobolium) with the fact that he dresses in rags for the ceremony 
(cf. McLynn 1996, 325). 

223 Cf. L.E. Tacoma, Fragile Hierarchies: The Urban Elites of third-century Roman Egypt. 
Mnemosyne Suppl. 271 (Leyden 2006). 

226 So righdy Schillinger 1979, 366f. 



268 CHAPTER FOUR 

the cult of the Mater Magna as a means of demonstrating that their 
private religious feelings by no means excluded a concern for the well- 
being of the state. 227 The donors of such altars (and so the preceding 
ceremony) represent a relatively wealthy minority of worshippers of the 
Mater Magna: most adherents were certainly of more modest means. 
From this perspective, we should think of the taurobolium as represent- 
ing an aspiration that the majority of individual worshippers of Mater 
Magna were never able to fulfil. 

Since Graillot's day it has been usual to distinguish two forms of 
the ritual: one on behalf of the emperor or a city the other private. 228 
In fact, however, 'public' taurobolia should be seen as one aspect of a 
more general desire among the patrons and members of the Metroac 
collegia to find appropriate means of honouring the imperial house. 229 
At any rate at the temple of the Mater Magna at Ostia, the fashion 
for taurobolia seems to work alongside and then to replace the intensive 
phase of offering dedications and imperial portraits associated with 
the Flavian to late-Antonine periods. 230 At the same time, the scope 
of the public institutions tends to widen: not only the entire imperial 
house but also the Senate, the XVwn, the ordo equester, the Roman army, 
the merchant-fleet may be included in the good wishes. 231 Conversely, 
'private' taurobolia have rightly been understood as in a sense collective 
events, even though the sacrifice was paid for by an individual and he 
or she was the primary beneficiary 232 The ritual was one means of 



22 ' Rutter published a very incomplete list of 1 15 items (some doubled) as an appendix 
to his work (1968, 243-9) with brief extracts; Duthoy reprinted 133 texts in full (1969, 
5-53), but these figures include many items from after AD 350. At least one 'public' 
taurobolium is known for: Antoninus Pius; M. Aurelius; Commodus, Septimius Severus; 
Clodius Albinus; Caracalla; Severus Alexander; Maximinus Thrax; Gordian III; Philip; 
Trebonianus Gallus and his son Volusianus; Valerian and Gallienus; Probus; Diocletian 
and Maximian. No 'public' taurobolia are yet known from the Germanies. 

228 Graillot 1912, 159; Rutter 1968, 236f 

229 The opposition between 'public' and 'private' in relation to Greek religion has 
rightly been questioned by the various contributors to V Dasen and M . Pierart (eds.), 
'iSia Ktxi Srjji.ooiq: Les cadres prices' et 'publiques' de la religion grecque antique. Kernos 
Suppl. 15 (Liege 2005); in a non-religious context: N. Bateman, Public Buildings: Some 
Contrasts, in B. Watson (ed.), Roman London: Recent Archaeological Work. JRA Suppl. 24 
(Portsmouth RI 1998) 47-57. 

230 Ostia is in fact the best example: Meiggs 1973, 360-5; Rieger 2004, 159-65; 
288-300; more generally, Schillinger 1979, 312-32. 

231 E.g. AE 1920: 92 = CIL XIV 40 = 4301; XIV 42 = ILS 526/4141 = CCCA 3: 
127 no. 405 (Ostia). 

232 CIL XIII 522 = ILAq Lectoure 13 = CCCA 5: 240: Severus Iulli fil. vires tauri quo propri(e) 
per tauropolium pub (lice) factum fecerat consacravit, is a private inscription that nevertheless 



THE RITUAL SYSTEMS 269 

committing the Metroac community to a particular ethical stance, one 
to which all could subscribe even without being themselves directly 
involved. 233 Likewise, I would argue, performing a taurobolium/criobolium 
to safeguard the social order by no means implies the absence of private 
concerns — we should not think of public and private here as mutually 
exclusive categories. A public taurobolium not only honoured the institu- 
tions of the state it also registered the higher status of the donor(s), since 
it required both permission from the local archigallus and the presence 
of civic priests as witnesses. 234 The withering of 'private' dedications 
from the mid-third century is probably a mirage caused by the general 
disappearance of private epigraphy of all kinds with the onset of the 
politico-military crisis, and thus not in itself at all indicative. 235 On the 
other hand, the social and economic disruption caused by the political 
uncertainties, to say nothing of events such as the deep raids by the 
Alamanni and the Franks, must have limited the funds available to the 
curiales and wealthy freedmen for such gestures. 

From this perspective, the fourth-century Roman taurobolium-dltars 
from the Vaticanum in Rome, almost all by members of the (by now 
largely side-lined) Roman elite, is not as remarkable as is often made 
out. 236 As McLynn remarks, "the taurobolium, it appears, had become 
the preserve of priests" (1996, 324). The late inscriptions focus upon 
the tauroboliated priest as an individual, not as a member of a col- 
lege; given that the real importance of the senatorial priestly colleges 
had shrunk to insignificance in the late fourth century, this was only 



insists the ritual was performed publics. This group of altars from Lactora/Lectoure in 
Aquitania also reveals that the priest Traianus Nundinius on a single day (8 Dec. 241) 
supervised one taurobolium celebrated by the entire ordo of the municipium (CIL XIII 
511= ILAq Lectoure 16) and eight for private individuals, mainly women (XIII 512—19 
= ILAq Lectoure 17—24). I think it likely that the date was chosen to mark the trib. pot. 
IV of Gordian III on 9 December. 

233 One or two texts note the fact that all the dendrophori and sacrati gathered to 
witness the sacrifice, e.g. AE 1897: 121 = CIL VIII 23401; AE 1898: 46 = CIL VIII 
23400; AE 1961: 201. 

234 CIL XII 1567 = LLS 4140 = AE 1982: 695 (Dea Augusta/Die, Gallia Narbon., 
AD 245) is a fine example of a private taurobolium being undertaken by a member 
of the curial class, accompanied by all the pomp of local officialdom. 

235 On the variable incidence of economic set-backs in this period, see recently R.P. 
Duncan-Jones, Economic change and the transition to Late Antiquity, in S. Swain and 
M.J. Edwards (eds.), Approaching Late Antiquity (Oxford 2004) 20-52. 

236 The most serviceable list is still that assembled in the Appendix to H. Bloch, A new 
Document of the last Pagan Revival in the West, HThR 38 (1945) 199-244 (of course 
without commitment to his overall analysis). For the dates on which these taurobolia were 
celebrated (spanning the period 12 March to 13 August), see Vermaseren 1977, 46. 



270 CHAPTER FOUR 

appropriate. These inscriptions recording private initiations read like a 
complete individual cursus honorum. The late-Roman taurobolium, we might 
urge, provided an occasion for the lavish public staging, and permanent 
memorialisation, of the personal commitment of these Roman senators 
to the continued performance of pagan cult. "An inscription mentioning 
a 'religious fact' is not necessarily a 'religious inscription'". 23 ' 

There has been some debate over the effects that the taurobolium was 
supposed to produce. It used to be assumed that its benefits, spiritual 
and material, were of limited duration, so that a renewal was necessary, 
apparently every twenty years. This period however is mentioned only 
by two late fourth- century inscriptions from Rome, and a passage of 
the Carmen contra paganos (1. 62). 238 (The rare references to (dies) natalis or 
natalicium in taurobolic contexts are irrelevant to this issue, since they 
seem to refer to actual, not spiritual, birthdays. 239 ) It may be appropriate 
here to recall that Lucius in Apuleius' Metamorphoses, Book 1 1 undergoes 
two initiations within a space of much less than twenty years and is 



237 J.F. Matthews, Symmachus and the oriental cults, JRS 63 (1973) 175-95 at 
182. 

238 CIL VI 504 = ILS 4153 (13 Aug. AD 376): vota...bis dem suscipit orbis; 512 = 
4154 (23 May AD 390): iterato, viginti annis expflejtis taurobolii sui (sic). VI 502 = ILS 4150 
(5 April AD 383) records a repetition but gives no period. Sempronia Eutocia and Ulat- 
tius Priscus in CIL V 6961-2 = ILS 4158-9 (Turin) apparently assumed the effects will 
be lasting: viribus aeterni(s) taurobolio (which I take to mean that the bull's testicles are to 
be preserved, e.g. in honey and spices, for ever, and that this has some connection with 
the expected duration of the benefit). McLynn has suggested that Sextilius Agesilaus 
Aedisius' claim that he has been in aeternum renatus (CIL VI 510 — ILS 4152, 13 Aug. 
AD 376) should be read not as "anti-Christian fanaticism but the over-eagerness of 
the parvenu" (1996, 327£). At any rate, it has generally, and rightly, been held that this 
is a purely personal or individual claim: Nilsson, GGR 2, 653; Duthoy 1969, 105-7; 
Sfameni Gasparro 1983, 206. Carmen: Martinez Maza 2000, 71ff A standard twenty 
years' duration was the usual older view: e.g. Hepding 1903, 198; Graillot 1912, 172. 

239 CIL II 5260 (Emerita): aram tauriboli sui natalici redditi; AE 1956: 255 = IRC Pacen. 
289 (Beja, Portugal): criobolati natali suo; CILXIll 573 (Bordeaux): natalici virib (us); 11352 

(Mediomatricum, Belgica): ara(m) t(auroboliatam?) ob naftjalicium ex iussu ; the idea that 

these were spiritual birthdays was suggested by M.-J. Lagrange, Melanges 1: Attis et 
le christianisme, III. Les mysteres et tauroboles, RevBib 28 (1919) 465ff; cf. Duthoy 
1969, 106ft It is thinkable that one's birthday (celebrated in antiquity on the first of 
the month in which the actual anniversary fell) was held to be a suitable occasion for 
such a sacrifice, held to promise special divine protection. It should however be noted 
that the Calendar of Filocalus (AD 354) uses the expression dies natalis inter alia to 
denote the day on which members of the house of Constantine attained imperial rank 
as Caesar or Augustus, e.g. Inscrlt XIII. 2 p. 251 (Constantine); p. 259 (Constantius), cf. 
Curran 2000, 224f. Such a usage must surely have emerged from a popular association 
between the literal 'birthday' and the metaphorical 'happy/fortunate day'. 



THE RITUAL SYSTEMS 271 

preparing for a third as the work closes. 240 There is therefore no inherent 
implausibility in the idea that the effects of the taurobolium, to the extent 
that it can be treated as an initiation, might have to be renewed. We 
can only say that there is no evidence for such a renewal before the final 
quarter of the fourth century; 241 that, though public renewal might well 
suit the interest of these late-Roman aristocrats in public display, only 
three such cases are known — that is, the great majority even of them 
either failed to renew, or died before they could do so; and there is only 
one extant case between AD 160 and 350 of the same person erecting 
a second taurobolic altar, namely Valeria Gemina at Lactora/Lectoure 
in AD 239 and 241 — and even here the earlier inscription seems to 
imply that she did not actually perform the taurobolium herself but was 
simply fulfilling the second part of the ceremony (vires accipere/consecrare) 
on behalf of someone else. 242 

The great majority of dedicators claim to have offered a taurobolium, 
no more and no less. That is, we must assume, they actually sacrificed 
such an animal. 243 A certain inflation, if that is what it is, begins to 
appear in the third century, when we occasionally find taurobolium et 
criobolium, both a bull and a ram, being offered. 244 A pair of altars from 



240 This may equally, however, be taken as a mark of Lucius' religious excess, or 
of the exclusiveness of initiation to a particular site or temple. Such exclusiveness is a 
typical strategy of rarefication. 

241 Assuming that CIL X 1596 = ILS 4271 (see n. 214 above), which includes the 
phrase iterata est, is not in the ordinary sense a taurobolic altar. 

242 CIL XIII 510 = ILAq Lect. 15: S.M.d Val(eria) Gemina vires escepit Eutychetis VIII Kal 
April(es) sacerdote Traianio Nundinio d.n. Gordiano et Aviola cos. (24 March 239). Noting that 
24 March was the dies sanguinis, Rutter (1968, 238), if I understand him correctly, sug- 
gested that Gemina offered the scrotum of a man named Eutyches who had castrated 
himself on that day. Why she should have done so, he does not say. I take it that this 
formula is an abbreviated allusion to that in CIL XIII 522 — ILAq Lectoure 13: vires tauri 
quo propri(e) per tauropolium pub (lice) factum fecerat consacravit (cited n. 232 above), and that 
Gemina consecrated the vires of a bull killed for (her husband) Eutyches, now deceased. 
The meaning "take over (from a predecessor), take up in turn" is well-attested for 
excipio (OLD sense 15; cf. sense 7: "take under one's care or protection"). It is at any 
rate noteworthy that Gemina does not state, as she does in her own taurobolic altar 
two and a half years later, that the victims were her own property, tauropolium accepit 
hosti(i)s suis: CIL XIII 518 = ILAq Lect. 23. 

213 At Dea Augusta/Die, Gallia Narb. the pontifex perpetuus of the civitas and his two 
daughters each gave a bull cum suis hostiis et apparam(entis) omnibus (see n. 234 above). 

244 CIL VI 505 = ILS 4143 (AD 295); 506 = 4144 (similar date); 508 = 4146 (AD 
319); VIII 4846 = ILAlg 1: 1983 (Severus Alexander); 23400-1 (AD 276/84 and 285/93); 
XII 1311 (first quarter IIP); 1745 (AD 209-11); AE 1924: 26 - ILGN 518; AE 1931: 
63; 1955: 49. Sometimes this is suggested simply by the imagery: e.g. CCCA 5: 133f. 
n. 386 (cited n. 216 above) and CIL XIII 1 1042 = ILS 9278 = CCCA 5: 146 no. 420, 
pis. CXLV-VI (Vesunna, III 1 '), which mention only a taurobolium, carry both a bull's 



272 CHAPTER FOUR 

Teate Marrucinorum/Chieti (Reg. IV) dedicated by the same man, 
Petronius Marcellus, a priest (probably of the Mater Magna), one on 
the occasion of his criobolium et (h)aemobolium, the other, in honour of 
Maximinus Thrax and his son Maximus (i.e. between early 236 and 
mid-April 238), on the occasion of his taurobolium, suggest that altars 
erected for a taurobolium et criobolium may in fact have commemorated 
two separate sacrifices, first the criobolium, then the taurobolium, perhaps 
undertaken several years apart. 245 Tempting though this is as a model, 
it is attested only in this one instance. 246 More than a century later, 



head with infulae and, less conspicuously, a ram's head. CIL XIII 1753 — ILS 4133 — 
CCCA 5: 137 no. 392, pi. CXXXV (Lugdunum, AD 194) carries a bull's head above 
the ram's head on one lateral face, but a ram's head above the bull on the other. 

245 CIL IX 3014-5 = ILS 4137-8 (= CIL II 180e* = II.5 p. 205 no. 48*) = CCCA 
4: 69 nos. 172f. This is the sole occurrence of the word (h)aemobolium in this context; 
though clearly of Greek origin, it is not recorded in LSJ 9 . We must assume that it is 
a reference to the collection of the animal's blood. Since the blood was collected in 
bowls at all sacrifices (and later made into blood-sausage etc.), the development of a 
special word implies that something slightly different was done with the blood at these 
Metroac sacrifices. There is no reason to suppose it was poured over the donor (cf. 
following note): that would be to show that Prudentius still lurks in the mind. 

2 "' A further complication involves the cernus/cerni. This vessel is mentioned in the 
singular and the plural in five inscriptions, four from North Africa, one from Rome 
(see Duthoy 1969, 99ff; AE 1897: 121 = 1898: 46 = CIL VIII 23401: there is only- 
one inscription here, not two). The earliest is from the period AD 169—77. Three texts 
from Mactar seem to classify the criobolium and taurobolium as sacra cernorum, i.e. as two 
members of the same class of rituals, "the ritual of the cerni", e.g. perfectis rit{a}e sacris 
cernorum crioboli et tauroboli (AE 1892: 18); perfectis ritfe sacris cernorum] et cri[oboli et tauroboli 
(AE 1959: 49). That from Utica however seems to treat the two as separate: cerno et 
criobolio de suo acceptis aram.. .fecerunt (AE 1955: 49; no taurobolium). At Rome the situ- 
ation seems again to be different, or is at any rate expressed in different terminology: 
taurobolium (et) criobol(ium) caerno perceptum per Fl. Antonium Eustochium sac(erdotem) (CIL VI 
497 — ILS 4146, AD 319). A Metroac cernus was certainly a vessel or container that 
could be carried in procession by a woman (e.g. CIL II 179, cernofora; EKtpvoipoprjcra 
ap. Clement of Alexandria, Protr. 2.15.3). Duthoy 1969, 73f argued that it could not 
possibly have contained the vires (he preferred blood); Rieger 2004, 166 believes on 
the contrary that they may have held ashes, horns — and/or vires. X Nicandr. Alexiph. 
217-21 asserts that Kepvoix; cpaoi xoix; |x\)axiKot)i; Kpaxfjpai;, e(p' (bv Xvxvovq xtGeaoiv: 
Hepding 1903, 9, cf. Sfameni Gasparro 1985, 68; 81. The Greek word KepvoQ normally 
denoted a ritual vessel with several sub-divisions, or small jars (cf. A. Bignasca, s.v. 
kernos, ThesCRA 5 [2005] 2b. VIA pp. 250-52) used for vegetable offerings and liquids; 
the scholion to Plato, Gorg. 497c however clearly thought they were the same as Jukvov, 
though this has been rejected in relation to the Eleusinian cult. Archaeology is thus 
of no help here. I incline to agree with Borgeaud (1998, 188f) that the texts from 
Mactar and Rome legitimate the inference that, at any rate in the cult of the Mater 
Magna in the High Principate, the vires of the sacrificial victims might be transported 
in cerni, whatever other uses these objects may have had (among the finds at the joint 
temple of Mater Magna and Isis in Mainz were large quantities of incinerated figs, 
nuts and other fruits, which had been offered to her: M. Witteyer, Verborgene Wun- 



THE RITUAL SYSTEMS 273 

this 'fused option' becomes quite regular among the wealthy senators 
in the Phrygianum, who could certainly have afforded to sacrifice both 
types of animal on the same occasion. Just a handful of individuals 
commemorate a criobolium and nothing else. 247 This might have been 
because they could not afford a bull; but it may equally be that they 
died before they could perform the taurobolium, or that the second altar 
was erected but has not survived. 248 We should therefore not jump to the 
conclusion that such donors were cheapskates. 249 It is however possible 
that some collective celebrations do attest financial constraints. An altar 
from Narbonne, for example, records the performance of a collective 
taurobolium at the command of the goddess ex stipe conlata, with the aid 
of a public subscription. 250 On these occasions, we may suppose, the 
members of the Metroac associations, with the approval of the city 
magistrates, went through the streets asking for alms. It is in fact likely 
that there was considerable local, and indeed individual, variation in the 
manner of handling taurobolia, which did not occur every day and for 
which the ritual procedures probably had to be re-invented, or creatively 
're-called', on each occasion. 251 The variety of different terms used in 



sche: Befunde antiken Schadenzaubers aus Mogontiacum-Mainz, in Broderson and 
Kropp 2004, 41-50 at 48). 

247 CIL II 5521 - ILS 4139 (Cordoba, AD 238: a woman; the man celebrated a 
taurobolium); XIV 41 = ILS 4135 = Rieger 2004, 295 MM 62B (Ostia, Commodan); 
VIII 8203 = ILS 4136 (pro salute imp. [Severus Alexander] and undertaken ex vaticinatione 
archigalli, so we have no reason to believe the donors were not well-connected); AE 1 96 1 : 
201 — CCCA 5: 40f. no. 114 (cerno et criobolio de suo acceptis, Q. Latinius Victor and his 
son, egregius filius, so Victor was presumably a primipilaris vel sim. The altar depicts a 
ram only). Perhaps also ILAfr. 201 (frag.). 

248 AE 1956: 255 = IRC Pacen. 289 (Pax Iulia/Beja, Portugal, cited n. 239 above) is 
an interesting case involving a father and son (duo Irenaei pater et filius) who held a criobo- 
lium to celebrate their birthdays. The date has been disputed: mid-II p : J. Encarnacao 
ap. IRC Pacen.; II— IIP: M. M. Alves Dias, Os cultos orientals em Pax Iulia, Lusitania, 
Memorias de Historia Antigua 5 (1981) 394; IV P : Duthoy 1969 no. 78. The first two are 
possible; Duthoy's date is not. I incline to think that Encarnacao is right. 

219 Since the arguments of Duthoy 1969, 61f, no one now believes in the old idea 
that the criobolium was specifically in honour of Attis. 

250 CIL XII 4321 = ILS 41 19: Matri deum taurobolium indictum iussu ipsius ex stipe conlata 
celebrarunt publice Narbon(enses) . Indictum here means 'proclaimed at a public meeting' 
rather than 'imposed'. 

251 The largest number of taurobolium inscriptions known from a single site, disre- 
garding the Vaticanum in Rome, is the twenty-two from Lactora/Lectoure in Aquitania 
(see n. 232 above). Disregarding the nine performed on a single day in December 241, 
we have evidence for thirteen taurobolia performed in the city over the century between 
AD 160 and, say, 260, i.e. 1.3 per decade (though in fact they seem to cluster in IIP'). 
Even granting massive epigraphic loss, these events cannot have been very common. 



274 CHAPTER FOUR 

the taurobolium-'mscriptions to refer to the sacrificial event surely had its 
correlate in the variety of actual procedures followed. 252 

Some have argued that, in view of the fact that the victim's scrotum 
was severed from the trunk and set apart, the taurobolium must have been 
a substitutive sacrifice. It is not always quite clear what is at stake here. 
One thought is that the severing of the animal's scrotum represented 
the subject's longing to satisfy the Mother's gelding-imperative while 
enabling him to attain the rank of priest without personal loss of the 
power of procreation. 253 

By implication, the 'true' worshippers of the Mater Magna were 
the galli, who actually had sacrificed their manhood. Maria Grazia 
Lancellotti has indeed argued that the taurobolium should be seen as an 
attempt to open the 'mystery' of the self-sacrifice of the galli to a wider 
public. "This can take place only through a 'revised reading' of the 
Gallus and the replacement of his real sacrifice with one that is sym- 
bolic". 2 ' 4 Everyone accepts that, whereas the ordinary galli had to geld 
themselves, the archigalli, as civic priests, must have been exempt from 
such a requirement, since among them are found full Roman citizens, 
to whom such practices were forbidden. 25,1 On this line of argument, 
the introduction of the taurobolium and the creation of the position of 
archigallus are intimately related. This would be the main achievement 
of the supposed second reform of the cult, under Antoninus Pius. On 
this view, it must have attempted to correct a number of problems that 
had emerged in the aftermath of the reform under Claudius. 216 



2:12 Known terms include taurobolium/viresfacere, perficere, accipere, excipere, movere, transferre 
(cf. Borgeaud 1998, 188). Note also expressions such as hoc taurobolio sacrum acceperunt 
{AE 1973: 579). 

253 E.g. Latte 1960, 354 n. 2; more recentiy, Turcan 1996b, 394. 

254 Lancellotti 2002, 114. She continues: "The mystai are the Galli/non-Galli, those 
who castrate themselves without castrating themselves". 

2 -" This is not a very strong argument, of course, since it assumes that such laws were 
intended to be enforced, rather than to have expressive force, and that that there was a 
high degree of internalisation of the law. Neither assumption appears very plausible. 

256 a. reform of the cult by Claudius was urged on the basis of Joh. Lydus, De mens. 
4.59 — Sanzi 2003, 310: Cybele no. 57, by J. Carcopino, La reforme romaine du culte 
de Cybele et d'Attis, 1: L'introduction officielle a Rome du culte d'Attis; II: Galles et 
archigalles, in idem, 1942, 49—75; 76—171. On this hypothesis, it was Claudius who 
introduced the archigallate to supervise the cult, thus reducing the galli to the role of 
exotic bit-players. As I mentioned above (n. 223) it was Lambrechts who first proposed 
a second reform under Antoninus Pius that established the archigallate. See now 
J. Alvar, El archigalato, in L. Hernandez and J. Alvar (eds.), Jerarquias religiosas y control 
social en el mundo antiguo. Adas del XXVII Congreso Internacional GIREA-ARYS IX (Valladolid 
2004) 453-458. 



THE RITUAL SYSTEMS 275 

One familiar objection to the theory that the taurobolium was essentially 
substitutive is that it was also performed by women, who of course had 
no testicles they could vicariously sacrifice. 2 " Perhaps two considerations 
are more telling. First, how could a substitutive rite, which only makes 
sense at the individual level, have become a largely public ceremo- 
nial, intended to maintain the welfare of the empire and the imperial 
house? If the taurobolium were primarily substitutive, how could there 
be collective rituals undertaken by entire communities? Second, does 
not the idea of substitution, as normally understood, assume that the 
galli somehow constituted the ideological or normative centre of the 
cult at Rome? A striking sight they may have been in processions, and 
fascinating they may have been to poets and satirists, but in the civic 
(epigraphic) manifestation of the cult, especially in Italy, it is representa- 
tive individuals and the colleges of the dendrophori and cannophori, with 
their close relation to the imperial house, that are far more visible. 258 It 
seems to me more plausible to argue that the primary function of the 
taurobolium was commemorative, closely analogous in fact to the annual 
felling of the pine-tree during the March festival and its transportation 
by the dendrophori into the temple area, a rite that recalled the act of 
Agdistis/Mater Magna cutting down the pine-tree under which Attis 
had castrated himself and died, and then bringing it into a cave, her 
dwelling-place. 259 According to Arnobius, Cybele gathered up Attis' 
bloody scrotum and washed it before burial. 260 This detail must surely 
be linked with the treatment of the scrota of the sacrificial animals 
alluded to in the epigraphy 261 An altar from Vesunna (Perigueux) in 



257 E.g. Rutter 1968, 23 7£, who, since he believes in the substitution-theory, finds 
himself forced to offer an extremely muddled 'explanation', that women were barred 
from performing public taurobolia (which he himself admits is not the case). Both Lan- 
cellotti 2002, 112 and Rieger 2004, 167 argue on the contrary that the very fact that 
women could perform the taurobolium proves its substitutive character. 

258 Meiggs 1973, 360-65; Schillinger 1979, 378: "Aber der Eindruck laBt sich doch 
nicht abweisen, daB, vereinfacht gesagt, in Italien mehr organisiert, in den Provinzen 
mehr gebetet wurde, oder anders ausgedruckt, daB eine Steinsetzung in Italien eher 
der Representation gait als der AuBerung personlicher Frommigkeit". 

259 Arnobius, Adv. nat. 5.7 with Sfameni Gasparro 1985, 40 with n. 66. 

21,0 Ibid.: sed abscisa quae fuerant Magna legit et f Mater deum, inicit his terram. Agdistis/ 
Cybele wraps the scrotum in Attis clothes; and violets appear on the grave. Both the 
clothes and the violets are directly linked by Arnobius to the ritual of the arbor intrat, 
cf. Mora 1994, 124f. 

261 Cf. Burkert, 1979, 120. As in the case of the galli discussed in the previous section, 
there is a good deal of uncertainty in e.g. Arnobius, Adv. nat. 5.12, about what sexual 
parts Attis and the others actually cut off: just the testicles or also the penis? 



276 CHAPTER FOUR 

the Musee du Perigord, where the bull is represented lying behind a 
statue-bust of Attis, so as to recall the shared loss of their vires, seems 
clearly to imply such a commemorative function for the taurobolium. 262 
But as the social status of Metroac worshippers rose, the preservation 
of the vires of sacrificial animals, as a metonym of the corpse of Attis, 
meant that the latter could be interpreted as a fitting metaphor for 
the salus of the individual and the Metroac community. From there it 
was but a small step to the custom of individuals donating animals in 
proprio nomine on behalf of the empire as a whole, with its constituent 
institutions. The vires thus came to evoke not so much to the castration 
of Attis as the preservation of his corpse. 

c. Initiation 

\pa ... toe dve^oiota e^oiao|iev Koci ra &veK^dA,r|ra 
eK^aWiaoiiev; 

Julian, Or. 5, 158d 

6i xiq av8pa, "EA,A,r|va r\ (3dp(3apov, |i\)eia6ai 
7iapa8oiri zic, maxtukov Tiva iru/ov imepfp'ufj Kd^iAei 
K<xi (ieyeSei, JtoAAa |a,ev opcovra irucmKa 9ed|iata, 

noXXav 8e ctKo-uovta toicutcov epeovcov 263 

Dio of Prusa, Or. 12.33 

In keeping with ideas about initiation into the pagan mysteries that 
were then widespread, Hepding claimed that initiates into the myster- 
ies of the Mater Magna were symbolically placed in a tomb. 264 One 
argument in favour of this idea is the word moriturus used by Firmicus 
Maternus in a passage relating to admission to the mysteries of the 
Mater Magna. 2fo Another is a passage of Sallustius that seems to affirm 



262 Inv. no. A3183 = CCCA 5: 146 no. 420, pi. CXLVI (left), cited n. 244; cf. Marco 
Simon 1997, 307. 

265 ". . .if you were to put a person, whether Greek or barbarian, in an overwhelm- 
ingly large and beautiful initiation-chamber, where he was to be initiated and witness 
many sights, and hear many voices, that are part of the experience..." (probably a 
composite image, not directly relevant to any specific cult). 

264 Hepding 1903, 196; Loisy 1932, 113. ' 

21)3 De errore 18.1. The passage reads: In quodam templo ut in interioribus partibus homo 
moriturus possit admitti, dicit. . . All modern editors since Hepding 1903, 49 read moriturus, 
which is the reading of cod. Vatic. Palat. 165, the sole ms. (which is for the most part 
carefully written, with a few corrections by a later hand, and some by the first editor, 
Mathias Flacius Illyricus). Lobeck 1829, 1: 24 suggested oraturus; Bursian (1856) and 
Halm (1867) preferred introiturus. 



THE RITUAL SYSTEMS 277 

that those who had simulated their deaths drank milk, as if they were 
newly reborn: 

Since the myth is so intimately related to the universe we imitate the 
latter in its order... and keep a festival therefore. First, as having like 
Attis fallen from heaven and consorting with the nymph, we are dejected 
and abstain from bread and all other rich and coarse food (for both are 
unsuited to the soul). Then comes the cutting of the tree and the fast, 
as though we were cutting off the further progress of generation (coomep 
Koti fi|i(ov d7iOK07ito|iev(ov tf|v 7t£pcmep(0 xr\q yeveaeaq 7ipoo8ov); after this 
we are fed on milk as though being reborn (cooTtep dvayevvr||iev(ov); that 
is followed by rejoicings and garlands and as it were a new ascent to the 
gods (Kod 7tpo<; xovq Seoix; oiov enavoSoq). 

Ilepi Geary 4.10, p. 8.16-25 Nock 

There are however strong objections to interpreting either passage in 
this sense. 266 The word moriturus in the passage from Firmicus Mater- 
nus is not a statement internal to the ritual he is alluding to: Who is 
the initiate? He is called 'One who is about to die', but part of his 
apologetic technique of signalling at every possible moment the folly 
of pagan religion in case one misses the point. 26 ' Moriturus is simply 
the deluded pagan worshipper, doomed to die in ignorance of Christ's 
message. 268 The passage of Sallustius is explicitly based not upon 
a ritual of personal initiation but upon the sequence of the March 
festival, particularly, it seems, the day of rest on 26 March before the 
lavatio on 27. The motif of milk seems to be tralatician: there is no 
mention of it in the symbolon cited by Clement and Firmicus Maternus 
(see below), 269 just as there is no evidence for the common assumption 



266 Sfameni Gasparro 1985, 80-82; the silence of Borgeaud 2004, 114-6 implies 
similar scepticism. 

267 For example: Audio Cinyram Cyprium templum amicae meretrici donasse — ei erat Venus 
nomen — , initiasse etiam Cypriae Veneri plurimos et vanis consecrationibus deputasse, statuisse 
etiam ut quicumque initiari vellet secreto veneris sibi tradito, assem unum mercedis nomine deae daret 
(10.1), where vanis, deputasse and secreto. . . tradito are all part of the same technique. See 
I. Opelt, Schimpfworter in der Apologie De errore profanarum religionum des Firmicus 
Maternus, Glotta 52 (1974) 114-26; D.M. Cosi, Firmico Materno e i misteri di Attis, 
Annali Fac. Lett. Pad. 2 (1977) 55— 81; J. Pepin, Reactions du christianisme latin a la 
soteriologie metroaque: Firmicus Maternus, Ambrosiaster, Saint Augustin, in Bianchi 
and Vermaseren 1982, 256—75; L.W. Barnard, L'intolleranza negli apologiste cristiani 
con speciale riguardo a Firmicus Maternus, Cristianesimo nella storia 11 (1990) 505—21. 

268 He goes on to claim: pestiferum veneni virus hausisti. . . cibum istum mors sequitur. . . alius est 
cibus qui salutem largitur (18.2). An exact parallel is to be found in 27. 1 : (Diabolus. . . disposuit 
ut...) perituros homines ex ligni imitatione deciperet. 

21,9 As far as I know there is only one item of evidence that may link milk with the 
cult of Cybele, an altar from Thessalonike mentioning a yaXaKxrupopoi;: L. Robert, 



278 CHAPTER FOUR 

that specific mystery initiations must have been held at the same time 
as the public festival. 270 In both cases, the cocraep-clause introduces a 
typically late, neo-platonising gloss, first the limit upon generation, then 
rebirth, defined by Julian in a parallel passage as 6cvo8ov tcov \|/v>xtt>v, 
the ascent of souls (Or. 5, 175b). 271 

The conclusion must be that we cannot describe initiation ritual in the 
cult of the Mater Magna in anything remotely approaching the detail 
available, thanks to Apuleius, Met. Book 1 1, for the cult of Isis. We can 
only say that there were such rituals, whatever their specific character. 272 
In all likelihood, such initiations developed only in the second century 
AD. 273 These general points seem to be confirmed by the occasional 
use of the cista mystica in Metroac contexts. 2 ' 4 The funeral monument 
of L. Valerius Fyrmus, priest of Isis and Mater Magna at Ostia, and 



Sur deux inscriptions grecques. I: Epitaphe de Gerasa. II: Inscription de Thessalo- 
nique, Melanges Bidez (Annuaire Inst. phil. hist, orientates de Bruxelles 2) 793-812 = Opera 
2: 988-1007. However, the association of this text with the cult of Cybele remains a 
mere inference (it has been convincingly attributed to the cult of Dionysus), and that 
is scarcely a strong basis for arguing in favour of Sallustius' claim. The use of milk in 
other initiatory contexts (Wyss 1914) cuts both ways. 

270 For the assumption, see e.g. Hepding 1903, 182—99; Vermaseren 1977, 116—19. 

271 Note that Julian here, despite his account of the successive rituals of the March 
festival (168c-69d) and extensive discussion of dietary restrictions (174a-78d), nowhere 
so much as mentions milk. 

272 Arnobius, for example, says that one of the sources he cites, Timotheus, claimed 
to have drawn his information about the myth ex reconditis antiquitatum libris et ex intimis 
eruta, quemadmodum ipse scribit insinuatque, mysteriis, from abstruse books of antiquities and, 
as he himself writes and claims, the secret teaching of the mysteries: Adv. nat. 5.5 = 
Sanzi 2003, 277f, Cybele no. 39.1. It is generally thought that Arnobius knew Timo- 
theus only indirectly, through Alexander Polyhistor (e.g. FGrH 273 F74); cf. Mora 1994, 
125-28. Whether mysteria here can pressed, I rather doubt: what 'mysteries' of Cybele, 
at least in our sense of the word, could Timotheus, assuming him to be a Hellenistic 
writer, possibly have known? A similar doubt applies to the claim by Harpokration 
that Neanthes of Cyzicus (presumably IIP) wrote up a (or the) Phrygian myth of Attis 
that was connected to the mysteries: iJAJcmKcx; 8e 6 X6yo<; (FGrH 84 F 37, presumably 
from the Teletai); cf. Sfameni Gasparro 1985, 52f For other references to 'mysteries' in 
connection with the cult of the Mater Magna (which however certainly sometimes refer 
simply to the galli) and should thus imply initiations, see Sfameni Gasparro 1985, 76. 
The most apparently promising reference, Julian's expression 8ta totx; n'oaxiKotx; Kal 
Kprjipioix; Geoiamx; (Or. 5, 169a), seems however simply to refer to non-public rituals 
during the March festival, whatever they may have been. 

273 Van Doren 1953; Borgeaud 2004, 114-6. 

2/1 Cf. Sfameni Gasparro 1985, 75f Cistae of course appear in a range of 'mystery' 
contexts, Eleusinian, Dionysiac and Isiac: X Aristoph. Lys. 642; Clement, Protrept. 
2.21.2 (Eleusis); LSCG no. 65 1. 29f. with Deshours 2006, 134 (Andania, i.e. Demeter); 
Theocritus 26.7-9; Catullus 64, 259f; IGUR 160, I A 25 (Dionysiac); cf. I. Krauskopf, 
s.v. Kiste, ThesCRA 5 (2005) 274f (for Isiac cases, see p. 317 below). Their contents 
certainly varied. 



THE RITUAL SYSTEMS 279 

especially the large cista that occupies Cybele's throne on a monument 
in London (provenance unknown), suggest that initiation was a familiar 
institution in the cult and not simply in the context of the galli} 15 

Some slightly more precise information seems to be offered by the 
symbolon preserved by Clement of Alexandria, cited in Chap. 2. Lb 
above: ek x\)\mavo\) ecpayov, ek Ki)|i|3dA,oi) etciov, EKEpvo(popr|oa, \)%b 
x6v naoxov traESw. 2 "' The Latin version offered by Firmicus Maternus 
provides an interesting variant: de tympano manducavi, de cymbalo bibi, el 
religionis secreta perdidici, while his Greek translation reads: ek xviinavov 
pEppcoiccx, ek K\)|j,(3dA,o\) 7t£7tcoKcx, yEyova [ivaxT^q "Axxeox;. 2 ' 7 These have 
been variously considered true variants or mere errors of transmission. 
In my view, they should be treated as complementary images. We might 
therefore suggest that 'conning the secrets', hierogamy and initiation 
to Attis are all ways of referring to the same reality 2 ' 8 

Despite the difficulty of drawing any clear conclusions about the 
content and scope of these rituals, such a perspective may allow us 
to interpret a good part of developed Metroac ritual by linking it to 
the myth. I would suggest that Attis' self-gelding acquires its full sense 
only if we allow that he enjoyed a sexual relation with Cybele, which 
was rudely interrupted by the proposed marriage to la. In the light 
of that, we should imagine a scenario in two parts, prepared for by 
prior instruction, purification and the successful fulfilment of certain 
trials or tests: the first would consist of a ritual banquet accompanied 
by the music of the typical Metroac type — 'flutes', cymbals, castanets, 
tambourines. I imagine this as a sort of eucharistic meal explicitly 



275 CttXIV 429 - CCCA 3: 152f. no. 466 - Meiggs 1973, 366 (possibly Isiac how- 
ever); CCCA 7: 1 If. no. 39, pis. xxvi— xxix (the ferculum is however being transported by 
a pair of galli). See also n. 195 above, on the cista depicted with Archigalli. 

276 "I have eaten from the tambourine, I have drunk from the cymbal, I have carried 
the sacred vessels, I have pushed past the curtain of the marriage chamber", Clement, 
Protr. 2.15.3 Stahelin, p. 24.1 If. Marcovich = Sanzi 2003, 266: Cybele no. 29.1. It must 
be admitted that, even if TidaTOi; here refers to a bed-curtain, in the absence of the 
ritual context we cannot tell whether it refers to a literal action (however understood) 
rather than evoking a specially important revelatory moment through the metaphor 
of the marriage-night. 

277 Maternus, De more 18.1 = Sanzi 2003, 267: Cybele 29.2. 

278 Vermaseren 1977, 1 16f. denies that the symbolon has anything to do with initiation 
(but see p. 118) and relates it to the March festival, and in particular the mesonyctium 
of Attis (the night 24-25 March). 



280 CHAPTER FOUR 

divided into solids and liquids (etpcxyov, etciov); or perhaps better, as a 
marriage-feast. 279 

After the meal the initiate, I suggest, proceeded to a sacred chamber 
where the hierogamy, the union with the goddess, was performed. 280 
It is quite uncertain whether we should suppose that this was a purely 
imaginary matter, or some sort of performance. 281 At any rate, from 
this point on, the worshipper was truly initiated and had access to the 
secrets of the cult. Nevertheless there remained the crime of incest, 
which could only be expiated by means the exemplary action of a few. 
To this end, the most devoted of the worshippers sacrificed their man- 
hood for the collective good, thus appeasing the castrating Mother. The 
ritual fulfilment of this altruistic act may have taken place more or less 
spontaneously on the dies sanguinis, 24 March. The reward was entry 
into the group of the galli. We may speculate that others confirmed 
their initiation through the taurobolium/criobolium, the commemoration 
of Attis' death confirming the link between Mater Magna and the 



279 Cf. Hepding 1903, 185-88; Grafflot 1912, 181; Vermaseren 1977, 119; Sfameni 
Gasparro 1985, 79f. 

280 I here follow the lead of Dieterich 1903, 126£; Graillot 1912, 182f; Pettazzoni 
1924/1997, 92; Burkert 1987, 98 n. 44; 107. For some counter-arguments, note e.g. 
Vermaseren 1977, 117£, who, relying on moriturus, believes it refers to a symbolic 
journey to the underworld. 

281 Cf. Nilsson, GGR 648£; Sfameni Gasparro 1985, 81. It seems unlikely that one 
can appeal for support here to the occasional occurrence of the term 0aXa|xr|ji6Xo<; 
in Metroac cult. The Hellenistic epic poet Rhianos devoted an epigram to Achrylis, 
QaXa]i.r\iiokoc, of Rhea = Cybele (AnthPal. 6. 173 = 67.1 Powell); Dioscorides uses the 
epithet of his figure Atys (AnthPal. 6. 220 Waltz = Gow-Page, 11. 1539-54, Dioscorides 
no. 16; see n. 102 above); cf. Pachis 1996, 201. Rather than there being a play here on 
the link between 0aXdur|/0dXa|XO<; (cave or lair; inner chamber, esp. marriage-chamber), 
we should perhaps take it that 0aXd|xr) in these contexts means a pectoral represent- 
ing the goddess in her shrine (as Gow explains the Dioscorides epigram). X Nicander, 
Alexiph. 6-8 explicidy says that 0aXdum were tojioi iepoi, unoyeioi, dvaKei|j.evoi xfj 'Pea, 
oitoi) eKxeuv6)j,evoi xa. ur|5ea KcaeTtGevTO oi x& 'A-rcei kou xr\ 'Pea XaTpewvxei;, "sacred 
subterranean places, consecrated to Rhea, where the servants of Attis and Rhea laid 
their severed genitals". Baslez 2004, 238, with n. 40, who wants to avoid all connection 
with self-gelding, naturally takes the word to mean "attache au service de la chambre", 
citing the existence of an analogous role in the cult of Bel at Palmyra — perhaps not 
a very convincing tack, since the speaker of Dioscorides no. 16 explicitly dedicates 
his ipiiv 9aXdur|v to the goddess, which he coud hardly do if it were actually part of 
the temple. Against that, S Eur. Phoen. 93 1 explains 6aXd|a.at as vessels used for safe- 
keeping objects; possibly therefore the equivalent of the cernus, cf. Sfameni Gasparro 
1985, 53. This interpretation is supported by e.g. the expression talamas suscepit c(h)rionis 
+ name in Duthoy 1969: 36 no. 75 = CCCA 5: 176 (Corduba, AD 234), and CIL II 
5521 = ILS 4139 = CCCA 5: 177 (AD 238). The word may therefore have acquired 
different connotations over time. 



THE RITUAL SYSTEMS 281 

worshipper. 282 By fulfilling such demands, the individual worshipper 
helped ensure that the goddess would continue to bestow her blessings 
both upon him- or herself, and upon the empire as a whole. 

Despite the poor evidence for initiation in the cult, I would argue 
that it was available to many individuals of both sexes. In order for a 
few to sacrifice their manhood as a sign of permanent commitment, 
there must have been a considerably larger pool of ordinary initiates. 
Although the dominant public image of them was negative, the galli 
must have been responsible for the spiritual oversight of the community; 
the archigallate, which did not demand gelding, was established in order 
to provide some social control over the cult as a whole. As we saw in 
relation to Portus, and doubtless elsewhere, a taurobolium undertaken for 
the public good ex vaticinatione archigalli, with the approval of the local 
archigallus, provided official dispensation from certain duties of tutelae. 2S3 
The office was thus afforded a degree of official recognition. From the 
mid-second century, the taurobolium/ 'criobolium became an option for the 
sub-curial class in some parts of the Latin-speaking west, as one means, 
alongside the dedication of imperial busts and statues, of expressing 
its commitment to the public weal. This sacrifice, alongside the March 
festival, and the Megalensia, became the public face of the cult of the 
Mater Magna — there is evidence that a procession of some sort was 
sometimes involved. 284 Taurobolia undertaken in one's capacity as a local 
dignitary 285 and the involvement of the curial class as a whole, indeed 
of entire cities, and provinces, represent the acme of this development. 
Insofar as it had a private face, we should see the sacrifice not as a 
substitutive offering but as a renewal of individual commitment through 
a distinctive ritual, both by individuals and by religious colleges such 
as the dendrophori and cannophori; also as a more or less expected under- 
taking if one became a priest of Mater Magna. 286 There is, however, 



282 So far as I know, all taurobolia/ 'criobolia are in honour of Mater Magna; Attis is 
never mentioned. 

283 See n. 222 above. The date is unknown, although it must ante-date Ulpian. 

281 E.g. CIL XII 1782 (p. 827): praeeunte Aelio Castrense sacerdote tibicine Albio Verino; XIII 
1753: praeeunte Aelio Castrense sacerdote, tibicine Fl. Restituto; 17 54: praeeunte Aelio Antho sacerdote 
sacerdotia Aemilia Secundilla tibicine Fl. Restituto apparatore Vireio Hermetione; AE 1910: 217 = 
1924: 26 = ILGJV 518: prae[euntib(us) sacerdotibus . . . 

285 E.g. CIL XII 385 = CCCA 5: 385 (Lugdunum); AE 1910: 217 = 1924: 26 = 
ILGJV 518: VIvir Aug(ustalis). 

286 E.g. CIL XII 1745: C. Valerius Ur[ba]nus sacerdos; 4322: Axia Cf. Paulina sacerd(os) 
(unless in fact abl.); AE 1892: 18: (1 Arellius Optatianus sacerdos eq. R.; AE 1898: 46: 
Minthonius Fortunatus sacerdos; AE 1955: 49: E Valerifus — Jtianus [sacerdos]; AE 1969/70: 



282 CHAPTER FOUR 

as we have seen in the previous section, clear evidence of shifts of 
(social) meaning over the long duration: there can be little doubt 
that in the late fourth-century its instrumental value to the remaining 
Roman elite was very different from what it had to been to a Q. Arel- 
lius Optatianus, eques Romanus at Mactar under Marcus Aurelius, or 
Q. Aquius Antonianus, pontifex perpetuus at Lugdunum under Commo- 
dus. 287 There must however have been many smaller differences over 
space and time that we simply cannot chart given the evidence available 
to us. 

e. The Megalensia and the March Festival of Attis 

scaena sonat, ludique vacant: spectate, Quirites 

Ovid, Fasti 4.187 

Festivals often involve some sort of dramatisation of the myth. In the 
case of the Mater Magna, the two main such festivals fell at moments 
that were highly significant for the cult. One, the spring festival, which 
at least in late antiquity lasted from 15—28 March, consisted of a 
symbolic presentation of Attis' passion. The other, the Megalensia, cel- 
ebrated shortiy afterwards, between 4-10 April, were founded in com- 
memoration of the arrival of Cybele at Rome in 204, and refounded by 
Augustus as the Ludi Matri deum Magnae Idaeae. 288 According to Cicero, 
these were in fact the only games with a non-Latin name. 289 In this 
case, the festival commemorated events important in the Republican 
reception of the cult, appropriately mythicised. 



1 19: Cornelius Gelastus sacerdos XVviralis M.M. I(daeae) Frygiae; AE 1994: 538: M. Rutilius 
Peculiaris sac(erdos) et libr(arius) public (us) XVvir(alis). These last examples suggest that 
sacerdos might subsume the archigallate. 

287 AE 1892: 18; CIL XII 1782 = CCCA 5: 369 (AD 184). 

288 CIL I 2 p. 314. The Fasti Antiates give the final day wrongly as 1 1 April. The 
suggestion by G.D. Hadzitis, The dates of the Megalesia, TAPhA 51 (1930) 165-74, 
based on the Calendar of Filocalus, that the games were actually held only on 4th and 
10th, has not found general acceptance: Ovid, Fasti 4.377-82 shows that at the time of 
Caesar they were held continuously over at least three days, cf. Herbert-Brown 1994, 
1 1 If.; Bernstein 1998, 201. Latte 1960, 436, rightly lists them as continuing for seven 
days. The first day commemorated the day on which the stone was brought into the 
temple of Victoria in 204 (Livy 29.14.14). The last commemorated the date in 191 
when the temple was consecrated, cf. J. Riipke, Fehler und Fehlinterpretationen in der 
Datierung des dies natalis des stadtromischen Mater-Magna-Tempels, J?PE 102 (1994) 
237-40, and p. 243 above. 

289 Cicero, de harusp. resp. 24. They remained such until the foundation of the ludi 
Adiabenici. 



THE RITUAL SYSTEMS 283 

Both festivals have received a great deal of modern commentary, 
and I do not want to discuss them here in any great detail. Both raise 
the issue of defining 'worshipper': processions within public festivals 
permitted a much wider range of participation, and a wider range of 
commitment, than we normally associate with the idea of 'oriental' 
cults. Much the earliest were the Megalensia, which should be thought 
of as but one of several competitions introduced in the half-century 
between the ludi Plebeii (220 BC) and the ludi Florenses (1 73). 290 They were 
organised by the curule aediles, and became, like the ludi Romani and 
others, an important part of aristocratic competition for social capital, 
since they provided an institutionalised opportunity for congiaria and 
extravagant display 291 The dramatic performances, introduced in 194, 
were staged in two different temporary wooden theatres in the heart 
of the city, at least one of them directly in front of the temple on the 
Palatine. 292 They included dances performed by the ballatores Cybelae, 
organised into a sodalitas. 293 Even in the late Republic, they seem to 
have included mimes or pantomimes of some version of the Phrygian 
myth. 294 The patrician character of these Games in the Republic has in 



290 See esp. Graillot 1912, 84-87; Lambrechts 1952a; P. Boyance, Cybele aux 
Megalesies, Latomus 13, 337-42 = 1972, 195-200; T.P. Wiseman, Cinna the Poet and other 
Roman Essays (Leicester 1974) 159— 69; Vermaseren 1977, 113— 25; H.H. Scullard, Roman 
Festivals (London 1981) 97-101; Turcan 1992a, 42-54; Bernstein 1998, 186-206. 

291 Scullard 1973, 24£; Gruen 1990, 24-6; Becher 1991, 162; K.-J. Holkeskamp, 
Rekonstruktion einer Republik. Historische Zeitschrift, Beih. 38 (Munich 2004) 93—105. This 
continued: as curule aedile in 65 BC, Caesar attracted notice through his financing of 
the Megalensia (Dio 37.8.1, cf. M. Gelzer, Caesar: der Politiker und Staatsmann 2 [Wiesbaden 
1960] 33). We might also think of the re-building of the Palatine temple after the fire 
of 111 BC (Pensabene 1996, 207), by one of the Caecilii Metelli (Ovid, Fasti 4, 348 
with Obsequens 39): Delmaticus (RE no. 91, cos. 119), his brother Numidicus (RE 
no. 97, cos. 109) or one of the sons of Macedonicus (RE nos. 77 or 84). This implies 
that it was of considerable popular interest, and therefore especially worthwhile for 
an aristocrat to repair. 

292 Livy 34.54.3 gives the date of the introduction of dramatic performances as 194; 
Valerius Antias frg. 40 Peter = FRH 1 5 F4 1 claims that they were first added in 191. 
They cannot have been the first such to be offered at Rome, since he himself elsewhere 
says that they were already part of the ludi Romani (FRH 15 F 38), cf. Bernstein 1998, 
193-5; Becher 1991, 162. On Cybele's gaze: Bernstein 1998, 204; the temple itself is 
briefly but competently described in Pensabene 1996. 

293 CIL VI 2265 = ILS 4179 = CCCA 3: 103 no. 361: sodales ballatore\s\ Cybelae, to 
one of their deceased members, T Flavius Chrysopaes. 

291 Taken as a.fabula praetexta by TRF p. 335 (but not printed as such in the text); 
cf. the anecdote about Octavian's alleged effeminacy recounted by Suetonius, Aug. 68; 
Becher 1991, 163; Nauta 2004, 61 Of. Nero's hymn to Attis may have been performed 
on the Palatine too (Dio 41.20.1—2). For a Christian view of the performances, see 



284 CHAPTER FOUR 

my view been much exaggerated (cf. Bernstein 1998, 199f), but there 
can be no doubt that special rules were attached to the performances: 
slaves were excluded, for one thing; and for the first time senators were 
separated from the rest of the audience, an early step towards the formal 
separation of the orders. 295 As I mentioned earlier, during the festival 
the aristocracy entertained one another to lavish banquets (mutitationes). 296 
The circus games, which are known only from the Principate, included 
a procession at which the sacred stone and a statue of Cybele with 
her lions were paraded through the Circus on fercula, and possibly also 
evocations of or allusions to the rites, such as the cista mystica mounted 
upon Cybele's throne. 29 ' The population threw roses into the street to 
pave her way with joyful colour. It was at the Megalensia too that the 
galli were traditionally permitted to collect the money tossed into the 
street by the spectators: aere atque argento sternunt iter omne viarum \ largifica 
stipe ditantes, (the people) strew her path with coins of copper and silver, 
bestowing upon her generous largesse. 298 

The March festival is a much more complex matter. 299 Although, as 
we have seen, many votive figurines of Attis have been discovered on 
the site of the Palatine temple, attesting to an intense popular concern 



Arnobius, Adv. nat. 4.38; cf. J.A. Hanson, Roman Theater-Temples (Princeton 1959) 13—16; 
Sfameni Gasparro 1985, 64. 

295 Livy 34.54.4-7; 36.36.3£; Cicero, de harusp. resp. 24; cf. J. Colin, Les senateurs 
et la Mere des Dieux aux Megalensia: Lucrece IV, 79, Athenaeum 32 (1954) 346—55; 
Beard, North, Price 1998: 1, 97. 

29,1 See n. 135 above. For the Hellenistic institution of the 5r||xo9oivia, to which this 
custom is presumably some sort of riposte, see P. Schmitt Pantel, La cite au banquet. 
Histoire des repas publics dans les cites grecques. CEFR 157 (Paris 1992) 255-424. 

297 Ovid, Fasti 4, 179-86 with Bernstein 1998, 201£; the details are confirmed by 
iconographical evidence, viz. a late-antique sarcophagus-lid from Aquileia (stone); 
another in S. Lorenzo fuori le mura, Rome (statue); see N. Himmelmann, Typologische 
Untersuchungen an romischen Sarkophagreliefs des 3. u. 4. Jhdts n.Chr. (Mainz 1973) 37-42 with 
figs. 58a, 56b. As is well-known, a large statue of the Mater Magna wearing a mural 
crown and riding a lion stood on the spina of the Circus facing the Aventine; this was 
the case, at any rate, in the imperial period from the time of Trajan: Tertullian, De 
spect. 8; Vermaseren 1977, 51— 4; J.H. Humphrey, Roman Circuses: Arenas for Chariot-Racing 
(London 1986) 273-77. 

298 Lucretius, Be rerum nat. 2, 626f; Ovid, Fasti 4.350-2; Ep. ex Pont. 1.1. 39f; cf. 
Cicero, de leg. 2.22; Dion. Hal., Ant. Rom. 2.19.4f. with Boyance 1972, 197f. 

299 The basic modern treatments, despite the problems connected with them, are 
Lambrechts 1952a; 1962, 42ff; 1967; cf. earlier Hepding 1903, 145ff; Graillot 1912, 
108—49; Carcopino 1942, 49—75; 76—171; important later accounts are Fishwick 1966; 
1967; Vermaseren 1977, 113—24; S. Fasce, Attis e il culto metroaco a Roma (Genua 
1978). In what follows, I take over Vermaseren's criticisms of Lambrechts' rather rigid 
model of reforms solely under Claudius and Antoninus Pius. 



THE RITUAL SYSTEMS 285 

with him, there is no mention of Attis as a god in literary sources of 
the Republic in the context of that site. 300 However, both Lucretius, de 
rer. nat. 2, 600-60 and Catullus 63, which is in 'gallic' galliambics and 
in form and subject related to popular mime, suggest there was already 
some knowledge of, and interest in, the Hellenistic account(s) in the 
late Republic. The galli were evidently familiar figures, implying that 
some sort of cult of Attis was practised. 301 On the basis of the conven- 
tional view of Augustus' distaste for foreign cults, it has been generally 
assumed, since the work of P. Lambrechts, that the basic form of the 
March festival emerged in the early Principate as a ritual of mourning 
for Attis and was formalised under Claudius. 302 A main argument was 
that, whereas the late-Republican and early-imperial calendars, such as 
the Fasti Antiates or the Fasti Maffeiani, list no public festival related to 
the Mater Magna towards the end of March, the Menologium Colotianum, 
of c. AD 50, does mention the Lavatio. 303 Both Peter Wiseman and the 
late Use Becher have however argued that the evidence of the Augus- 
tan poets by no means supports the idea that Augustus was hostile to 
the Phrygian cult as a whole. 304 There is thus no particular reason to 
date the emergence of the early stages of the March festival as late as 
Claudius. As we shall see, there is actually some archaeological evidence, 
quite apart from Ovid, Fasti 4, 179-372, to suggest that the goddess' 
bath was shifted away from the Palatine temple, and by implication 
moved to the Almo, already in the Augustan period. 



3110 See n. 137 above; also Varro, Menipp. frg. 150, p. 27 Astbury: cum illoc venio, video 

gallorum fiequentiam in templo dei As I pointed out above, the position of Lambrechts 

1952b and 1962 on this point is unsustainable; see the discussion by Thomas 1984, 
1503f 

301 Bremmer 2004, 558£; cf. Harder 2004, 587£; M. Fantuzzi and R. Hunter, Muse 
e modelli (Rome and Bari 2002) 558f.; for fabulae togatae called Megalensia, see Afranius, 
frg. 219f. Ribbeck = 222f Daviault; Atta frg. lOf = 1 If. Daviault. Maecenas' piece 
on the myth of Attis, like Catullus', was in galliambics: Maecenas, frgs. 4f. Baehrens 
= 4 Lunderstedt = 5-6 Courtney (cf. n. 187 above). Courtney's second, anonymous, 
fragment: rutilos recide crines habitumque cape mi, urges a follower of Cybele, presumably 
a gallus, to abandon that life. 

302 Joh. Lydus, de mens. 4.59: ttiv eopTrjv (i.e. of the Arbor intrat on 22 March) KXav8io<; 
6 PacuXetx; KaxeoxrioaTO. 

50:5 CIL I 2 p. 280f. (with p. 314) = VI 2305 = ILS 8745 = Inscrlt XIII.2, 47 (p. 287). 
Given the nature of this unofficial calendar, it cannot be concluded that there were 
no other Metroac celebrations at this date. The traditional Quinquatrus (or -id) were 
celebrated 19-23 March. 

304 Wiseman 1984; Becher 1991; Herbert-Brown 1994, 114f; Bernstein 1998, 
196f. 



286 CHAPTER FOUR 

The March festival was no sort of ready-made construction according 
to an already-developed programme of ideas. Its scope grew over at 
least two centuries, probably more. It is in fact only in the mid-fourth 
century that we have reliable evidence, from the calendar of Filocalus 
(AD 354), for a fully-developed sequence between Canna intrat on 15th 
and the Initium Caiani on 28th. 30S This list, representing the final point 
of a long evolution whose details remain obscure, is as follows: 



Id. Mart. 


(15 March) 


Canna intrat 


XI K. Apr. 


(22 March) 


Arbor intrat 


IX K. Apr. 


(24 March) 


Sanguem 


VIII K. Apr. 


(25 March) 


Hilaria 


VII K. Apr. 


(26 March) 


Requietio 


VI K. Apr. 


(27 March) 


Lavatio.™ 6 



Rather than go through the list in its calendrical order, as is often 
still done, it seems preferable to emphasise here the extent to which 
the March festival was an historical construction subject to continual 
alteration and innovation, both in its elements and in the significance 
attached to them. 307 

The earliest festival seems to have been the Lavatio, which is mentioned 
already by Ovid as a popular occasion. 308 Since Romanelli discovered 
a sizeable basin (3.05m X 3.65m X 1.70m) lined with blocks of Grotta 
Oscura tufa in front of the Palatine temple, and framed by its steps, 
where the statue (or the stone) was apparently washed under the Repub- 
lic, it is likely that the procession to the Almo was an Augustan innova- 
tion, that is, Ovid is alluding to something recent (yet another case of 
an Augustan invented tradition). 309 The sacred stone was taken from the 
Palatine temple to the Porta Capena in a procession that included the 



305 CIL I 2 p. 260 - Inscrlt XIII. 2, 42 (p. 243). It is doubtful whether the initium Caiani 
has in fact anything to do with the March festival, rather than with the taurobolium (the 
Gaianum was part of, or near, the Phrygianum): it is usually ignored (cf. Degrassi, 
p. 433 s.v. Mar. 28). Lambrechts inclined to think that the old Megaknsia were replaced 
by the new festival, but this seems impossible (they are still listed in the Calendar of 
Filocalus). 

306 See Fishwick 1966, 193f. 

307 I here extend the point made after Lambrechts 1952b by both Fishwick 1966; 
1967 and Vermaseren 1977, 13-24. 

308 Ovid, Fasti 4.339-42; cf. Mommsen ap. CIL I 2 p. 314. As Borgeaud 2004, 64 
points out, the passage fuses the parade of 4th April with the later Lavatio in the Almo. 
For the other texts on the Lavatio, see Wissowa 1912, 319 nn. 4 and 7. 

309 Romanelli 1963, 303; Pensabene 1978; Becher 1991. 



THE RITUAL SYSTEMS 287 

XVviri s.f. ? ' w Its course alluded to, without at all precisely following, the 
account of the goddess' reception in 204 BC associated with the gens 
Claudia: from the Porta Capena in the Servian Wall it proceeded down 
the Via Appia to the evidently small, rather insignificant, temple of the 
goddess on the right bank of the Almo, now the Aquataccio, a rivulet 
that flows sluggishly into the Tiber and from 7 BC marked the south- 
ern boundary of the Augustan Regio I (Porta Capena) of the city 311 
Here the statue and the sacred iron utensils were bathed xcov Opnycov 
voufi), according to the Phrygian custom, by a priest, robed in red; 312 
the XVviri prayed that the goddess might return safely to the Palatine; 
and the return journey, lit by torches, was the scene of gay rejoicing: 
exululant comites, furiosaque tibia flatm; etferiunt molles taurea terga manus? u ' 
If the Lavatio were indeed an Augustan innovation, we may assume 
that it was introduced at about the time of the Augustan rebuilding of 
the Palatine temple in AD 3 after a fire attended by a miracle, 314 and 
was officially intended to register the links between the Great Mother, 
Troy, the Sibylline oracles (now kept in the temple of Mars Ultor) and 
Roman invincibility, however it was understood by the galli or the mass 
of the population. 315 

So far as is known, the Lavatio never had anything to do with Attis, 
who is not mentioned in any of the relevant sources. Nevertheless 
there are clear signs of a shift in its significance. At latest in AD 140, 



510 This of course was the college that supervised the Sibylline Books, which had 
originally commanded that the goddess be brought to Rome (Livy 29.10.4f.). The only 
evidence that the stone was set into a silver frame or decoration is Prudentius, Perist. 
10.157, and is therefore suspect. 

311 For the course of the Servian walls and the approximate boundaries of the 
Augustan regiones, cf Kolb 1995, 404 fig. 65. 

512 According to Ammianus Marcellinus, RG 23.3.7, the carpentum, the covered wag- 
gon, was also washed. Although Ammianus was sufficiently interested in the cult to 
write an entire excursus in his lost account of Commodus on the arrival of Cybele 
at Rome in 204 BC (22.9.5E), this detail is probably a misunderstanding (he does say 
perhibetur); cf. R.L. Rike, Apex omnium: Religion in the Res gestae of Ammianus (Berkeley 
1987) 49f 

313 Ovid, Fasti 4.341f. (description of the supposed events of 204 BC; the molles 
manus that beat the drums are of course those of the galli); Lucan, Bellum civ. 1.600: 
et lotam parvo revocant Almone Cybeben; Valerius Flaccus, Arg. 8.239—42; Martial 3.47.2; 
Arrian, Tact. p. 33.4 Roos. 

311 Ovid, Fasti 4.437E; Augustus, RG 19; Valerius Maximus, Mem. 1.8.1 1. 

315 It has been supposed that the washing was supposed to cleanse the stone of the 
blood of the galli (e.g. Vermaseren 1977, 121 but this is surely a false inference from 
Valerius Flaccus, Arg. 8.241 : quis modo tarn saevos adytis fluxisse cruores cogitet? No one doubts 
that the vires (mystic or real) were kept in the penetrate. 



288 CHAPTER FOUR 

and no doubt somewhat earlier, a procession of the collegium hastifero- 
rum was introduced, the spear-carriers of Ma Bellona. 316 By the fourth 
century, and probably much earlier (i.e. the second century), it had 
become attached, at least loosely, to a festival of Attis, celebrated on 
22 and 24 March, and known, at least in the fourth century, as Arbor 
intrat and Sangufinjem. 317 We must assume that these were the focus of 
mourning at this time for the death of Attis. The new defixiones from 
the temple of Isis and Magna Mater in Mainz, whose concern with 
the self-laceration of the galli and the bellonarii (the ecstatic followers 
of Ma Bellona, whom we know to have been closely associated with 
the cult of the Mater Magna) I have already mentioned, suggest the 
hypothesis that it may have been this sight, of the self-inflicted pain, 
the still-warm blood spattered on the ground to grow cold, that stimu- 
lated wide-spread interest in the cult of Attis, and eventually induced 
Claudius, if it was he, to mark this celebration officially 318 

Arbor intrat centred upon the death of Attis beneath a pine-tree. 319 In 
its developed form, the members of the collegium dendroph(or)orum felled 



316 The earliest inscription to mention the hastiferi, from Ostia, is dated 140: AE 
1967: 74; the temple of Bellona there was constructed late in the reign of Hadrian. 
The repeated allusions in the Mainz defixiones, dated AD 70— c. 130) to galli and bellonarii 
imply that the connection was older. Herodian 1.10.5—7 shows that by the time of 
Commodus the emperor himself could be expected to be present at the procession in 
Rome, cf. Fishwick 1967, 148—50, 154—7 (I agree with Fishwick that the reference is 
to the return from the Almo, rather than the Hilaria, since there is no reason whatever 
to think that the latter was considered important enough to warrant the presence of 
the emperor). Schillinger 1979, 390f. takes the emperor's presence, perhaps rightly, as 
specific to Commodus himself, and thus not to be generalised. 

317 On the date of Arbor intrat: Julian, Or. 5.168b7 claims that the ritual was performed 
on the day of the spring equinox; the Menologium however specifically gives the date of 
the equinox in P as 25 March, later the day of the Hilaria; and it appears from 171c 
that Julian had other reasons for wanting to link Attis with the equinox. Macrobius, Sat. 
1.21.10 (in connection with the Hilaria) also dates the equinox to 25 March; Sallustius, 
KEpl 6ewv 4, p. 8.26f. Nock, carefully says 'about the time of . . . about the time of (wepi). 

318 E.g. Blansdorf 2008, no. 18 1. 3f. ita uti galli bellonarive absciderunt concideruntve . . . : 
no. 16 1. 5f: quomodo] galli se secant et praecidunt vir[i]lia sua, sic ilk. . .; 1. 9fi: quomodo galli, 
bellonari, magal[i] sibi sanguin[em]ferventem fundunt, frigid[us] ad terrain venit, sic... I. 1 If.: 

quern] admodum de eis gallo[r]u[m, majgalorum, bellonfariorum — ] spectat sic et illi membra 

m[ed]ullae extabescant. These magali are otherwise unknown, but they must be a similar 
group of ecstatic followers attached to the cult. The name is evidently derived from 
a fairly uncommon Greek appellation of Cybele, MriTrip |j,eydXr| (e.g. CCCA 2, 131 
nos. 430f, from Lebadeia). Bellonarii is evidently a popular or slang term, like many 
words in -arii; for the assocation of Ma Bellona, Mater Magna and the hastiferi, see 
Fishwick 1967, 152-7. 

319 It is usually thought that a fragmentary relief found in Bordeaux in 1838 showing 
four muscular men manhandling a log depicts the Arbor intrat: CCCA 5, 145 no. 416; 



THE RITUAL SYSTEMS 289 

the tree before dawn in the sacred grove, decorated it with woollen 
ribbons and wreaths of violets, and then carried it into the Palatine 
temple, where it lay (no doubt for days, or even weeks) withering. This 
ceremony was accompanied by scenes of violent grief. 320 Although 
the earliest epigraphic evidence dates from 9 April AD 79, the Greek 
name of the dendrophori clearly implies a much older institution. 321 The 
considerable number of votive plaques representing the dying god pre- 
sumably attests to the intensity of worshippers' self-identification with 
his fate. 322 After an interval of one day, spent in mourning, Sangu(in)em 
commemorated the blood that flowed from Attis' self-inflicted wound: 
the galli and bellonarii processed through the streets lacerating their arms 
with knives and double-axes and lashing themselves with whips knotted 
with astragali.™ It was generally believed that this was the day on which 



Vermaseren 1977 fig. 73. I think this very doubtful: the new Mainz defixio (see next 
note) specifically mentions the withering of the tree, which must therefore still have 
had its needles on; the great log shown in the Bordeaux relief is quite inappropriate. 
Rather, the Attis-relief from Glanum (CCCA 5, 1 1 7fi no. 344, pi. CXIX = Vermaseren 
1977 pi. 63; see PI. 20) shows the sort of thing required: a small, young pine behind 
the grand mythical pine of the main scene surely alludes to the tree of Arbor intrat. 

320 See Suet. Otho 8: die quo cultores deum matris lamentari etplangere incipiunt. All the other 
sources are fourth-century or later; the fullest is Arnobius, Adv. nat. 5.16 = Sanzi 2003, 
283: Cybele no. 39.4; see also Julian, Or. 5, 168c; Firmicus Maternus, De errore 27.1; 
Carmen c. paganos 108; Joh. Lydus, De mens. 4.59. The detail about the tree withering 
is taken from an image in the Mainz defixiones: ita uti arbor siccabit se in sancto, sic et illi 
siccet...: Blansdorf 2008 no. 18 1. 7. 

321 CIL X 7 (Regium Iulium), in honour of a number of women who had made gifts 
to the college. In Ostia, as elsewhere, although in the second century the dendrophori 
seem to have been an exclusively religious association, they later came to be closely 
connected to other workers in wood (fabri tignarii) and the centonarii (fire-brigade), e.g. 
AE 1987: 198f; AE 1957: 80; CIL V 5128 = AE 1993: 800, cf. Rieger 2004, 168f 
They continued to be a religious group into the late-Roman period, however, and 
were for this reason suppressed in AD 415 by Honorius (CTheod. 16.10.20). Although 
the patrons of the collegium dendroph(or)orum are of relatively high status, the known 
magistrates were socially insignificant: of those at Ostia, for example, none is known 
to have taken part in the municipal administration at any level (though one or two 
were Augustales), and none is known to have been a magistrate of any other collegium 
(which is another typical mark of good social standing): H.L. Royden, The Magistrates 
of the Roman Professional Collegia in Italy (I-III AD) (Pisa 1988) 57-9. 

322 See conveniently Vermaseren 1966, 30—44. 

323 The date is given by Tertullian, apol. 25.5 and HA Claud. 4.2 {dies sanguinis); I take 
it that Tertullian's archigallus is merely a high-sounding synonym of gallus. There is no 
other good source for the date (Lactantius, inst.div. 1.21.161 gives none). For what it is 
worth, the HA claims that many members of the Senate were present in the temple on 
this day in AD 268; unfortunately it cannot be correct, since Claudius did not become 
emperor until Sept. /Oct. of that year. This is not the only indication of parti pris in 
relation to Claudius' accession, cf. R. Syme, Emperors and Biography: Studies in the Historia 
Augusta (Oxford 1971) 204-6. 



290 CHAPTER FOUR 

individuals might geld themselves. The symbolic aim however was to 
express the asymmetry of the bond between worshipper and divinity; 
the shedding of blood associates the worshipper with a sacrificial ani- 
mal, the suffering body (these self-lacerations were undertaken every 
year) is presented as a palimpsestic text of devotion. 324 Probably on the 
evening of 24 March Attis was 'buried', which may be the occasion of 
the mesonyctium, the all-night vigil mentioned earlier. 323 

The earliest epigraphic evidence for the cannofori falls in the reign of 
Antoninus Pius, i.e. the second third of the second century. Most of it 
is later. 326 Fishwick rightly argued that Canna intrat formed a major part 
of the re-structuring of the March festival at this period: it was perhaps 
intended to extend the quasi-narrative sequence by creating an allusion 
to Attis being found by Cybele among the reeds of the river Gallus. 32 ' 
It is at any rate obvious that the ritual was calqued upon Arbor intrat: 
the cannofori cut bundles of reeds and ceremoniously carried them to the 
temple. There are again signs that the significance of the ritual altered 
over time: Joh. Lydus associates it with fertility, whereas Julian seems 
to link it with Attis' emasculation. 328 Perhaps of greater importance is 
the date: at a stroke, apparently, the length of the mourning and fast- 
ing period was extended by six days prior to Arbor intrat, perhaps once 
again under the influence of popular pressure. 329 On the other hand, 
we should recall that the festival of Anna Perenna, with its drunken 
merriment and amorous play fell on the day immediately preceding 
Arbor intrat (21 March), so that we should not exaggerate the extent of 



324 Nevetheless I cannot understand Fishvvick's claim (1967, 144) that the dies sanguinis 
was 'the climactic point of the festival of Magna Mater'. That was the Lavatio. 

325 CCCA 5: 133f. n. 386, pi. CXXXI, cited n. 216 above. 

326 Fishwick 1966, 637; Rieger 2004, 170: the earliest texts are again from Ostia: 
OIL XIV 40 = 4301 = Rieger MM62; and 117 = Rieger MM47. 

32/ The only significant text is Julian, Or. 5 165b (Attis exposed at birth); Sallustius, 
rcepi 0£(ov 4, p. 6.28 and 8.5f Nock does not state that he was abandoned, merely that 
Cybele saw him by the river. There is an allusion to Attis hidden among the reeds of 
the river Gallus on the modius of M. Modius Maxximus from Ostia (see n. 169 above), 
with Attis in his Phrygian cap above between two reeds, and the bearded god below 
(Vermaseren 1977, 114; Beard 1994, 5f.). 

328 Joh. Lydus, mens. 4.49 (the mention of a six-year-old bull in a procession led by 
the apxiepeix; [= ? archigallus?] however suggests some muddle with the taurobolium; but 
it may be correct); Julian, Or. 5, 168d. 

329 The Quinquatrus was evidently no longer of much importance, if indeed it 
ever had been. 



THE RITUAL SYSTEMS 291 

general mourning at Rome — or even awareness that the March festival 
was being celebrated. 330 

The last festival to be introduced into the sequence was apparently 
the Hilaria (25 March). 331 There is no reference to it even as late as AD 
136. 332 But in the mid-fourth century, perhaps under Christian pressure, 
it becomes an important theme in more sophisticated accounts of the 
March festival, such as that of Julian. 333 Here again, however, there are 
signs of varying interpretation. 334 Whereas Julian and his friend Sal- 
lustius refer the festival to the ascension of the individual worshipper's 
soul, the fifth-century neo-Platonist Isidorus son of Theodotes appar- 
ently thought it was a matter of being saved from Hades (e^ "ASov> 
ocoxripiav). 333 These views are at least related; Firmicus Maternus' 
account (though it does not explicitly name the Hilaria) focuses upon 
Attis as a figure for the vegetative cycle, in particular of grain: it 
dies annually at the hands of the reaper, is sown again and annually 
comes back to life. 336 On the other hand, there is a certain amount of 



"° On the festival of Anna Perenna in the light of the new defixiones from her 
well beneath Piazza Euclide, see now T. Wiseman, The Cult-Site of Anna Perenna: 
Documentation, Visualization, Imagination, in L. Haselberger and J. Humphrey (eds), 
Imaging Ancient Rome: Documentation — Visualization — Imagination. JRA S61 (Portsmouth RI 
2006) 51-62. 

331 And doubtless the Requietio, about which nothing is known, but which was 
evidently merely a bridge to the Lavatio. On the whole confused question, see CIL F 
p. 312; Hepding 1903, 107-72; GraiUot 1912, 131-6; Lambrechts 1952a; 1967. Fishwick 
1967 and Vermaseren 1977, 119—23 are especially important and helpful discussions. 
Vermaseren's criticisms of the arguments of Lambrechts 1952a and 1967 are wholly 
justified (cf. Sfameni Gasparro 1985, 58 n. 137), even though he comes to much the 
same conclusion in practice. 

332 -phis ; s jhg probable date of Arrian's Tactica, where he gives a brief account of 
the March festival (see n. 313 above). I cannot accept Sfameni Gasparro 's argument 
(1985, 57) that Val. Flaccus, Arg. 8.239-42 refers to expressions of joy, taken as an 
antecedent of the Hilaria, directly after the mourning — the passage explicitly refers to 
the Lavatio. She also cites (57 n. 131) HA Sev. Alex. 37.6 as evidence for the existence 
of the Hilaria in the third century. The parts of this life that derive from the Kaiserge- 
schichte are indeed relatively reliable (Barnes 1970, 34—8), but other sections, and this 
chapter on his eating habits in particular, are full of bogus material: Barnes 1970, 33. 
HA Aurel 1.1 is no less bogus. 

333 Or. 5, 169d; Sallustius, Tiepi 6ewv 4, p. 8.25f. Nock; Macrobius, Sat. 1.21.10; 
Damascius, Vit. Isidort : §131, p. 176 Zintzen. 

334 See also Cosi 1986, 107ffi; LanceUotti 2002, 135-42. 

335 There seems no reason to believe that these views go back, say, to Porphyry. 

336 De errore 3.2. In his commentary (3.3f.), Maternus himself studiously avoids talk- 
ing about joy, emphasising only the mourning for the death; but it is implicit in this 
cycle-theory: vitam rursus quod iacta semina annuls vicibus reconduntur. As LanceUotti points 
out, however, we must beware of ancient 'exegeses of exegeses' — especially, but not 
only, Christian ones (2002, 138). 



292 CHAPTER FOUR 

iconographic evidence, in the form of bronze and terracotta statuettes, 
hard to date but certainly from the High Principate at latest, and some 
surely (late) Hellenistic (found at Myrrhina), that represent Attis danc- 
ing. 337 Vermaseren argued that these represent Attis celebrating his 
return to life. Since this must have been an aspect of the mystery-teach- 
ing, however, and is treated as such by Julian and Sallustius, it seems 
improbable that it would have been so widely represented in vernacular 
art. In the same way, I see no reason to link images of winged Attis 
with the theme of resurrection. We should probably understand the 
late emergence of the Hilaria as a named festival as a response to neo- 
platonist and even Christian emphasis upon ascent/salvation, albeit on 
the basis of an element in the mystery experience that must be a good 
deal earlier. 338 It is difficult to go any further. 

By far the most salient rituals connected with the cult of the Mater 
Magna at Rome were public ones. That is why we dispose of relatively 
large amounts of information concerning them. The negative side is 
that, insofar as initiation is concerned, almost no information survives, 
and what there is has been filtered through tendentious Christian writ- 
ers. We may nevertheless argue that it was precisely the experience of 
repeated participation in specific, structured ritual performance, above 
all procession, mourning, fasting, and rejoicing — that is: fairly consistent 
formal features, legitimising notions and memory 339 — that encouraged 
the formation of what we might call a Metroac habitus, which in turn 
formed the basis for the desire on the part of a much smaller number 
to construct a dynamic emotional and ethical relationship with Attis 
through initiation, whatever quite we are to understand by that word 
here. 340 The best evidence we have for the importance of forming 
such a personal, quasi-mystical, relationship are the plaques represent- 
ing Attis' self-gelding and death: in the one from Glanum illustrated 



337 Vermaseren 1966, 47—56. 

338 Cf. Bremmer 2002, 50-55. 

339 Granted the point made by A. Chaniotis that 'one cannot celebrate the same 
ritual twice' (2006, 234). In relation to the self-wounding of the followers of Isis at 
the Ida, Firmicus Maternus remarks: veterum vulnerum resecant cicatrices ut annuis luctibus 
in animis eorum funestae ac miserandae necis exitium renascatur (de errore 2.3). The body too is 
a palimpsest. 

340 As I mentioned earlier (see n. 294), it is probable that under the Principate there 
were performances re-telling at least some of the story of Attis at the Megalensia too, 
thus providing both a link between the two festivals and a modicum of information 
within the population of Rome. Otherwise however the ludi are better understood in 
the context of Republican Roman religion. 



THE RITUAL SYSTEMS 293 

here (PL 20), as commonly, the intended relationship is implied by the 
god's frontal gaze out of the picture-plane towards the spectator. The 
institutional expression of that commitment is surely membership in 
the collegia of dendrofori, cannofori, and hastiferi, which are quite widely 
recorded in Italy and the western empire (Schillinger 1979, 312—32). 
The taurobolium on the other hand represents the integration of the pub- 
lic role of Metroac cult with the desire for a special form of individual 
emotional commitment among the curial and sub-curial class. Under 
some limited and unusual circumstances, the same desire might reach 
expression in the longing for a radical separation from the world of 
common social reciprocity in the act of self-gelding — itself considered 
a form of 'mystery'. But this extreme choice must always have been 
attractive only to a few. 

4. Rituals in the Egyptian Cults 

For lack of evidence, it is virtually impossible to reconstruct the ritu- 
als peculiar to the cult of Serapis. 341 We are not justified in simply 
transferring to his cult the fairly extensive information we possess for 
the cult of Isis. For we do not in fact know what sort of relation there 
subsisted in daily practice between the two: although we speak of 
'the cult of Isis and Serapis', they often appear separately both in the 
iconography and in inscriptions. We can affirm neither that they were 
separate nor that they were worshipped without much discrimination 
between them. 342 We can only say that they do often appear together 
in votive inscriptions throughout the Principate, and that they often 
appear conjointly, especially on lamps. 343 We cannot however specify 
how this coexistence was figured in practice. We do not know whether 
the adherents of the one god also worshipped the other, how far offerings 
and votives were directed to them jointly, or even if temples at those 
sites where there were no separate shrines were normally dedicated to 



341 The Menologium Colotianum (see n. 303) records the Serapia for April; it seems 
still to have been celebrated in the third and fourth centuries, since the Calendar of 
Filocalus dates it 25 April (CIL P p. 259 - RICIS 501/0221). It is however otherwise 
completely unknown. 

342 So righdy Dunand 1973, 1: 79f 
313 Malaise 1972a, 198. 



294 CHAPTER FOUR 

them jointly. 344 My own inclination is to think that they did normally 
function as a pair; 345 but there are some problems with such a view. 

We know that some temples at least were dedicated to one of them 
alone: Pausanias for example tells us that on the way up to Acrocorinth 
there were two shrines to Isis, one called Pelagia, the other Egyptian; 
and two to Serapis, one of them 'in Canopus' (2.4.6). 346 This case 
highlights another point that has received relatively little attention, 
that differentiation such as this corresponds to the amount of stress 
laid on the Egyptian or Greek nature or aspect of the divinity in each 
case. 347 Thus Isis may be worshipped specifically as Aegyptia, Memphitis 
or Bubastis, Serapis as ev Kcxvomcp or NeiXdycoyoq; on the other hand, 
Serapis may be ev NaimcxKTcp or ev Ilopxcp while Isis may receive 
Graeco-Roman epithets such as Capitolina, Augusta, euploia, Pharia or 
pelagia, which latter are found along the Mediterranean littoral, or on 
islands, where this particular cult was most intensively practised. 348 
The case at Acrocorinth also suggests that each temple must have had 



341 This seems often to have been the case in Italy (Malaise 1972a, 135); in the Iberian 
peninsula there are joint temples at Ampurias and Cartagena (Alvar 1993c), but not 
in the case of the Isea at Italica and Claudia Baelo: J. Alvar and E. Muniz, Les cultes 
egyptiens dans les provinces romaines d'Hispanie, in Bricault 2004b, 69—94. 

345 Many scholars, such as Merkelbach 1995, 86f. §15 If.: 'als heilige Familie abge- 
bildet', have simply assumed that Isis, Serapis and Harpocrates really did function as 
a sort of 'holy family', pointing to cases such as the bronze plaque in the Antiquarium 
at Pompeii, which shows Sarapis and Isis standing right and left of a base or pedestal 
on which Harpocrates stands holding a cornucopiae: see Tran tarn Tinh 1983, 107 
no. IB 2; 1990 no. 167; Merkelbach 1995, pi. 68; or the relief from the Via della 
Conciliazione now in the Museo Capitolino, showing Sarapis enthroned in the cen- 
tre, flanked by Harpocrates and Isis on the spectator's right and Demeter to the left: 
Helbig 1 2 no. 1185 = Merkelbach 1995, 608 fig. 138 = Bottini 2006, 254f no. 62. In 
fact the triad Serapis-Isis-Anubis is epigraphically much more common (see RICIS 2: 
774f), but nobody dreams of calling this collocation a 'holy family'. 

346 Cf the temple in Sicily, celeberrimum etfrequentissum but otherwise unknown, where 
Verres erected his own statues (Cicero, IlVerr. 2.160). It was dedicated to Serapis 
alone. 

34 ' On the hellenisation of Isis, see Malaise 2000, stressing the greater depth and 
intensity of the 'hellenising' evidence. 

348 "Sanctuaries of Isis or the Great Gods of Samothrace help map the networks 
of redistribution in the period of their popularity": Horden and Purcell 2000, 442. As 
a matter of fact, however, as I have mentioned already (nn. 1 14f above), the epithet 
pelagia occurs as such only five times, though there are several images on relief and 
on specially-formed lamps (Bruneau 1961; 1963; 1974; Blanchaud 1984; Tran tarn 
Tinh 1990, 794; Bricault 2006a, 84—90). There is however a fine glass gem now in 
the Kunsthistorisches Museum, Vienna (inv. AS IX a 991), showing Isis Fortuna with 
the Pharos 'lighthouse': Mostra hide no. IV 239 = Ausstellung Liebighaus 683f no. 286 
(M. Pfrommer). On the Egyptian epithets, see now Bricault 2005a. 



THE RITUAL SYSTEMS 295 

its own particular character and that the rituals observed must have 
differed from one to the other: given that they were so close together, 
there would have been no point in reproducing the same rituals in all 
four. They must have arranged things differently between one another, 
and similarly with the temple at Cenchreae, the Saronic Gulf (eastern) 
port of Corinth, no distance from Acrocorinth, which is the setting of 
Apuleius, Met. Book ll. 349 

Serapis is mentioned just once in this narrative, in the context of the 
flamboyant description of the procession to celebrate the Isidis navigium: 
ibant et dicati magno Serapi tibicines, qui per obliquum calamum adaurem porrectum 
dexteramfamiliarem templi deique modulumfrequentabant.™ The fact that these 
flautists had nothing to do with Isis seems clear from the words dicati 
magno Serapi and familiarem templi deique modulum: this was music specific 
to Serapis and his temple in Cenchreae. It might be urged that the later 
references to Osiris are in reality to Serapis. 351 But it is obvious that in 
Apuleius' view the two divinities are quite different: the initiation into 
the mysteries of Osiris has nothing to do with Serapis, and is further- 
more quite separate from that undertaken for Isis. All this inclines me 
to think we should not rush to conclusions but assume that they were 
sometimes worshipped together or jointly, sometimes separately. 

Be that as it may, there can be no question that our information 
about ritual practice in relation to the Isiac group of divinities (includ- 
ing Anubis, Harpocrates, Osiris and Horus), here called 'the Egyptian 
cults', greatly exceeds that for either of the other two cults here dis- 
cussed. This also explains why they have been the focus of much more 
academic discussion. On the other hand, we would do well to reflect 



349 For the arguments in favour of identifying the apsidal building on the western 
side of the south mole as the Iseum, see Scranton, Shaw and Ibrahim 1978, 53—78; 
Egelhaaf-Gaiser 2000, 167-71; Scranton and Ramage 1967, 138-52 refer to the 
partly-submerged building simply as 'the sanctuary'. A brief account of the temple at 
Kenchreai, and a plan, in Bommas 2005b, 319—21. 

350 "pip ers dedicated to Serapis who, on transverse pipes held close to the right 
ear, repeated the traditional melody of the god and his temple": Apuleius, Met. 11.9, 
tr. J.A. Hanson. 

551 As in Lactantius, Inst. div. 1.21.22: "This is the Osiris whom people call Serapis 
or Serapides" (tr. A. Bowen). This is simply a suggestio falsi, as is plain from e.g. the self- 
predication of Karpokrates from Chalkis in Euboea: KapKOKpaTn, ZapdinSi, ctKoaii; 
"IoiSo.;, 'Oaeipi5i ejiriKocp: RICIS 104/0206 1. 1 = Totti 1985, 15 no. 6. Lactantius 
himself notices that something is wrong, and hastily explains: "Names usually get 
changed when the dead are deified". 



296 CHAPTER FOUR 

on the possible costs of such heavy reliance upon a single, contested, 
text, Apuleius, Met. Book 11. 

a. Festivals 

The cult of Isis as celebrated outside Egypt represented an extreme 
simplification of the colourful surfeit of festivals, pan-Egyptian, regional 
and highly local, relating to Osiris, Horus, Isis and Anubis in the pre- 
Ptolemaic and Ptolemaic period. 352 The Graeco-Roman reception 
observed just two major festivals a year, 353 only one of which, the Isea, 
was important in Egypt itself. 354 This itself attests to the extent to 
which what the Graeco-Roman world received as 'Egyptian religion' 
was a highly selective, and constantly re-interpreted, version of the 
original. 305 

The first, which I have already mentioned, is the Isidis navigium, whose 
Greek name was ra neyaXa nXoioMpeam. 33 '' At least in the later empire, 
it was held on 5 March. 35 ' Here Isis was made known as the protec- 
tor of navigation, whose annual season formally opened on that day, 



352 Cf. POxy 1380 1. 202 = Totti 1985, 62ff. no. 20: 'IoeTa raaeih;] itoXeaiv ei<; 
tov [ajtavjra %povo[v Katjeopcjrioai; (many letters uncertain), with the list of sites in 
11. 1-74; Bilabel 1929; Dunand 1973, 1: 1 10-62; 207-44. The use of the word 'IoeTa 
is itself a mark of the Hellenisation of such events in the Ptolemaic period. 

353 Cf. Merkelbach 1963, 39—41 (the link with earlier Egyptian festivals is totally 
unconvincing, however, cf. Perpillou-Thomas 1993); 1995, 147—86 (very adventurous). 

354 There were three major pan-Egyptian Isiac-Osirian festivals (on the dates, it 
should be remembered that the traditional Egyptian (Sothic) calendar was mobile, every 
month slowly altering its location in the seasons of the solar year, so that the actual 
date varied; after 199 BC, intermittent attempts were made to celebrate festivals on a 
fixed day, appointed with reference to the solar calendar, cf. Bickerman 1980, 40f): that 
of the onset of the Nile flood, nominally on 1 Thoth of the Sothic year (= 25 Epeiph 
= 19 July Julian); the harvest-festival, on 1 Pachon of the Alexandrian calendar (26 
April); and the Khoiak festival (Dunand 1973, 1 : 217; F. Daumas, s.v. Choiakfeste, LAg 
1 [1975] 958-60). Only the last, it seems, was commonly observed outside Egypt. 

355 This was nothing new: there is plenty of evidence (e.g. Herodotus 2.61) that the 
Greeks in Egypt from an early point selected and adapted Egyptian religious institu- 
tions and practices to suit themselves. 

356 RICIS 114/0703 (Byzantium); Apuleius, Met. 11.17. The Latin name is attested 
only by the Menologium Colotianum {CIL I 2 p. 280f = VI 2305 = ILS 8745 = Inscrlt 
XIIL2, 17 - RICIS 501/0219) and the Calendar of Filocalus {CIL I 2 p. 2562 = RICIS 
501/0221). 

357 Lactantius, Inst. div. 1.11.21; CIL I 2 p. 260 (Filocalus); Joh. Lydus, Mens. 4.45; 
I can find no evidence in support of Merkelbach's claim (1995, 157 §291) that the 
date at one time varied according to the full-moon, for which anyway he provides no 
citation. Wissowa thought it likely that it was introduced at Rome at much the same 
time as the Isia, i.e. under Caius (1912, 354). 



THE RITUAL SYSTEMS 297 

when prayers were offered on behalf of those who plied it that the sea 
remain calm, and yield up its harvest once more. 338 As Isis pelagia, she 
was proclaimed mistress of the sea, inventor of navigation and protector 
of sailors. 359 The goddess was thus linked to the encouragement and 
protection of trade (Apuleius, Met. 11.5): 



Diem qui dies ex ista nocte nascetur aeterna mihi nuncupavit religio, quo sedatis 
hibernis tempestatibus et lenitis maris procellosis fluctibus, navigabili iampelago rudem 
dedicantes carinam primitias commeatus libant me sacerdotes. :m 

At the same time, the navigium recalled Isis' voyage in search of her 
husband/brother Osiris. It was this that allowed her worshippers to 
address their prayers to her as Isis Fortuna, and thus all the more 
able to intervene in aid of sailors, merchants and travellers of all sorts 
who needed to cross the sea. The institutionalisation of this festival 
is an excellent illustration of the extent to which the Egyptian cults 
managed to integrate themselves into the ideological superstructure 
of the Graeco-Roman city: an essentially foreign deity, from a culture 
regarded at best with ambivalence, gradually turning into a protector 



358 Although the idea of a Mediterranean sailing season was conventional (e.g. Vege- 
tius, epit. rei mil. 4.39, dating it from 27 May until 14 September), there seems always 
to have been a good deal of sailing activity even in autumn and winter: Horden and 
Purcell 2000, 142f Aelius Aristides, for example, went by sea from Rome to Miletus 
and Smyrna in mid-winter AD 144 after his intestines swelled up and he had to return 
home (Or. 48 — Sacred Tales 2.5.60-9; on the date see Swain 1996, 265). 

359 The theme occurs twice in the self-predications: Cyme etc. (Totti 1985, 1 no. 1 = 
RICIS 302/0204 etc.) §15 and 39; cf. Andros (Totti 1985, 5 no. 2 = RICIS 202/1801) 
1. 34f. On the navigium in general, see Witt 1971, 165ff; Malaise 1972a, 217-21; Dunand 
1973, 3: 223-30; Griffiths 1978, 31-46; Turcan 1992a, 114f. Stern 1953, 225 rightly 
rejected the claim by D. Levi, Antioch Mosaic Pavements, 1 (Princeton and London 1947) 
37; 164f that one of these mosaics represents the navigium, but considered that it 
might be the subject of the well-known Isiac procession now in the Vatican (Helbig 1 
1, no. 491 = Cumont 1929, pi. VIII. 1 = Merkelbach 1995, 615 fig. 145) (see PL 22 
here). Malaise 1972a, 221 suggested that the festival may have been represented on 
the walls of the Isiac temples of Pompeii, Ostia, Herculaneum and on the Aventine, 
cf. Tran tam Tinh 1964, lOOf On the whole subject, see now Bricault 2006b. There 
is an interesting marble boat at Beneventum that originally represented Isis standing 
in it, of whom only one foot survives (Malaise 1972b, 300 Beneventum no. 15, with 
pi. 57/8 — Ausstellung Liebighaus p. 402 fig. 6, I a ). 

360 " r phe J^ w hich will be the day born from this night has been proclaimed mine 
by everlasting religious observance: on that day, when the winter's tempests are lulled 
and the ocean's storm-blown waves are calmed, my priests dedicate an untried keel to 
the now navigable sea and consecrate it as the first fruits of voyaging", tr. Hanson. 



298 CHAPTER FOUR 

of sea-going communication. 361 This is an important example of the 
assimilative capacity of Roman religion: it has been plausibly suggested, 
for example, that images of Isis with cornucopiae and steering-oar 
evoke the annona brought annually from Alexandria to Rome. 362 At the 
same time, the festival is also an expression of the cult's alterity: though 
its theme was so central, it was staged not by the civic magistrates 
but by the priesthood and the members of the relevant collegia and 
sodalicia, the pastophori, iepa(popov, EimcxKoi, 'Iaicxaxca, Zapoc7uoccn;ca, 
Zi)Vocvoi)|3icxaT(xi and all the other groups occasionally mentioned in 
the epigraphy. This evident support for the common good affirmed the 
commitment to the public weal of a group otherwise unable to take a 
leading part in civic rituals. The Egyptian cults thus demonstrated their 
unequivocal commitment to social order and solidarity, their superficial 
alterity notwithstanding 363 

The festival consisted essentially of a procession down to the shore- 
line to mark the first sailing of the new year. 364 Apuleius first describes 
the procession, in all its colourful variety (11.7—11). After an interlude 
in which Lucius is turned back into a man (11.12-15) the procession 
reaches the shore: 

There, after the images of the gods had been set in their proper places, 
the chief priest (summus sacerdos) consecrated a ship, which was constructed 
with fine craftsmanship and decorated all over with marvellous Egyptian 
pictures. He took a lighted torch, an egg, and sulphur, uttered prayers 
of great solemnity and reverent lips, and purified the ship thoroughly, 
naming it and dedicating it to the goddess. The gleaming sail of this holy 
barque bore an inscription woven in letters of gold, whose text renewed 
the prayer for prosperous navigation during the new sailing season. Now 



361 There is as yet no evidence at Rome or in Italy for the celebration of the Isidis 
navigium already in the Roman Republic: for example, the supposed Isis euploia or pelagia 
on a late-III p types from Syracuse is probably an Aphrodite: Sfameni Gasparro 1993 
[1995], 90—2. However the festival surely came from the Hellenistic Greek world; the 
obvious source would be Delos, where the Serapea and the Iseum were sacked in 88 
BC, and only the Iseum rebuilt (Dunand 1973, 2: 98f). 

362 Bricault 2000a, 141f.; 2006b, 83-84; 171f; cf. earlier Malaise 1972a, 181; Bru- 
neau 1974, 381. 

363 For one approach to the political instrumentalisation of Isiac festivals, see Hidalgo 
1995. 

set r pj le procession carrying images of gods from one temple, to another, or around 
the countryside, was one of the typical ritual modes of Egyptian religion: Dunand 1973, 
1: 210-12; U. Rosler-Koller, s.v. Gdtterbesuch, LdA 2: 669-71; Frankfurter 1998, 38f; 
52—58. The Isidis navigium however represents a significant variation upon this pattern 
(Bricault 2006b, 134-150). 



THE RITUAL SYSTEMS 299 

rose the mast, a round pine, high and resplendent, visible from afar off 
with its conspicuous masthead. The stern curved in a goose-neck and 
flashed lightning from its coating of gold-leaf . . . Then all the people, 
worshippers and uninitiated alike (tarn religiosi quam profani), outdid one 
another in loading the ship with baskets heaped with spices and similar 
offerings, and on the waves they poured libations of grain-mash made 
with milk. When the ship was laden with generous gifts and auspicious 
sacrifices, it was untied from its anchor-ropes and offered to the sea, as 
a mild breeze rose specially for her. After her course had taken her so 
far that we could no longer clearly make her out, the bearers of the 
sacred objects (sacrorum geruli) took up again what each had brought and 
joyfully set out on the way back to the shrine, preserving the order and 
fine appearance of their procession. 

Met. 11.16, tr. Hanson (adapted) 

Once the cortege has returned to the temple, the ritual proceedings are 
brought to a close. The chief priest (Tcpo^triq, tepoypanuaxevx;), the 
tepcxtpopot and those already initiated enter the penetralia and replace the 
divine statues in their proper places. The priest then summons the pas- 
tophori (who did not count as priests), and presumably the remainder of 
the participants in the procession, and reads prayers, evidently in Latin, 
for the safety of the emperor, the state, all mariners and their vessels. 365 
He then performs a short ceremony in Greek formally to inaugurate 
the new sailing-season. The audience breaks out into acclamations, 
and its members push forward to lay greenery at the feet of the god- 
dess' silver statue and kiss her feet (exosculatis vestigiis deae): m They then 
disperse back home, filled with a sense of joy (Apuleius, Met. 1 1.17). 367 
Although the scene is set at Cenchreae, we may assume that the festival 
at Ostia or Portus, where an important group of Alexandrian shippers 
connected with the annona was based, was broadly similar. 368 No doubt 
at Rome the procession simply went down to the Tiber: Isis was also 
mistress of rivers. 



365 Cf. n. 28 above. 

366 For similar acclamations after a festival, see Chaniotis 2006, 233f., citing LSCG 51 
(Iobakchoi), with A. Schafer, Raumnutzung und Raumwahrnehmung im Vereinslokal 
der Iobakchen von Athen, in Egelhaaf-Gaiser and Schafer 2002, 173-220 at 188. 

367 Cf. Chaniotis 2006, though, focusing on sacred laws, he does not cite this case. 

368 Cf. e.g. RICIS 503/f207, 1215, 1216, 1217; Meiggs 1973, 279; 387. A Roman 
festival of (Isis) Pharia is recorded for April by the Menologium Colotianum (RICIS 
501/0219), but not entered in the Calendar of Filocalus. How it differed from the 
Isidis navigium is quite unknown. 



300 CHAPTER FOUR 

The origins of the Isidis navigium are obscure. At Philae in Upper 
Egypt there was a festival which involved transportation on a litter of 
the 'boat of Isis', decorated with her image fore and aft. 369 This boat was 
both a cult object and in some sense identified with the goddess herself; 
and it was closely associated with the celestial voyage of the Sun-god 
Re. Both at Philae and at Abydos, there were other boats that brought 
Osiris to his sister/wife but were also in some sense identified with Isis 
(Dunand 1973, 1: 215). These may be linked with the 'navigation of 
Osiris' known from the Canopus decree and a well-known inscription 
from Thessalonike, a poem by Damaios dated c. 120 BC. 370 The most 
important antecedent of the Graeco-Roman festival is implied by an 
inscription from Eretria of the first century BC recording a dedication to 
Serapis, Isis, Osiris, Anubis and Harpokrates by T Septomios Ptolemaios 
Damos and his wife or sister Septomia Antiochis, voa>ocpxr|aavT£c;. 3/1 
It seems likely that when Isis' solar boat and/or that of Osiris were 
received in the Aegean, via trading contacts with Alexandria and/ or 
the Delta cities, a creative misinterpretation occurred that transformed 
them into a novel competitor for the Great Gods, an opportunity taken 
up and extended by the colporteurs. 372 

The other annual festival took place between 28 October and 
3 November. At least according to the Calendar of Filocalus, it bore 
the overall name Isia. 37i As Mommsen already argued, it was based on 
the Egyptian festival of Khoiak (which was celebrated in December) 
shifted, presumably because of the exigencies of the Alexandrian civil 
calendar, to 19 Hathyr =15 November Julian. 374 It was doubtless first 



369 E dem. Heidelberg. 736 ap. W. Speigelberg, ^eitschrift fiir Agyptische Sprache u. Alter- 
tumskunde 1917, 33f.; Dunand 1973, 1: 211-4. 

370 OGIS I no. 56 1. 41 and 54; IG X.2: 108 11. 2f.= Totti no. 72 = RICIS 
113/1506. 

371 IG XII Suppl. 565 = RICIS 104/0111, cf. Dunand 1973, 2: 26. The stele is 
decorated with two crowns. 

372 Analogous to the creative misunderstanding that turned St. Leonard, a saint of 
prisoners, into the protector of livestock, esp. horses. 

373 GIL I 2 p. 274, 276 = RICIS 501/0221 (giving 28-31 Nov. and 1 Dec. only); Joh. 
Lydus, de mens. 4.48 gives the last day as 3 November, the date of the Hilaria = Heuresis of 
the Menologium (Stern 1968). The painted calendar of S. Maria Maggiore (early IV P ), 
published in 1972, gives Luct(us) Isidis (so Stern) for 28 October, but the festival might 
evidently begin somewhat earlier, between 25-28: Stern 1973; RICIS 501/0220. 

374 CIL I 2 p. 336, cf. Wissowa 1912, 353f; on the contradictory Egyptian evidence 
concerning the date, see Parker 1950, 41; Griffiths 1970, 312; 448f; Dunand 1973, 1: 
227f; Perpillou-Thomas 1993, 94-100. Of the two current spellings, Athyr (Egyptolo- 
gists) and Hathyr (papyrologists), I choose the latter. 



THE RITUAL SYSTEMS 301 

introduced at Rome in the late Republic (cf. Ovid, Met. 9.693f.), but 
may have been first established in the calendar around AD 25 and 
28; 375 if so, it must have preceded Caius' hypothetical (re-)building of 
the Iseum Campense. 3 ' 6 The festival was essentially a ritual staging of 
the myth of Isis and Osiris. Such stagings cannot have been anything 
unusual, if we can go by the anecdote told by Suetonius concerning 
the eve of the murder of Caius, that is, the night of 23/4 January AD 
41: parabatur et in noctem spectaculum, quo argumenta inferorum per Aegyptios 
et Aithiopas explicarentur (Caius 57.4). 377 Such performances evidently 
reproduced in different forms the central events of a narrative that 
purported to explain the order of the cosmos and the system of Isiac 
beliefs (see Chap. 2.1. a). The first part of the festival was devoted to 
the passion (luctus) of Isis, her desperate search for her husband/brother. 
The worshippers shaved their heads, beat their breasts and slashed their 
arms as they made as if to hunt for Osiris, until they finally 'found' 
him. 3 ' 8 The poem by Damaios from Thessalonike, mentioned earlier, 
clearly alludes to the joy of Isis when the body of Osiris is found. 379 
According to Plutarch, whose account is not readily reconcilable with 
the Roman reports but seems to be partly based on the ritual followed 
at Dendera in Upper Egypt, the priests went down to the sea, took a 
golden casket from the sacred kioxt| that apparently contained the head 
of Osiris, poured some water into it, and the onlookers cried evprnoivoi) 
%ox> 'OoipiSoq. A statuette of earth, water and spices was then made, 



575 Stern 1968; Lembke 1994, 67; 89f.; Takacs 1995, 90. Seneca, Apocol. [Ludus] 13, 
written c. AD 54, assumes that his readers will be perfectly familiar with the Isiasts' 
cry £'upT|Ka|j.ev, cruyxaipcouEV (slightly misquoted by t, Juv. 8.29f). It is thought that the 
ueA,avr|<p6poi played a major part in these mourning rituals. 

376 Besides supplying a fanciful reconstruction of the Iseum (p. 149 fig. 5), due to 
his apparent ignorance of the standard work on the topic (Lembke 1996), Cavalieri 
2005, 148—54 seems to me to overemphasise its role as a carrier of Domitianic dynastic 
propaganda in pursuance of his general argument. 

577 Josephus, 47 19.1.12-14, §84f, 94 and 101 indicates that the murder took place 
after Caius had left the theatre on the Palatine. 

378 Seneca, de superstitione (ap. Aug., De civ. Dei 6.10.46 Dombart-Kalb): cum perditio 
eius inventioque fingatur; idem, ap. Servius, adAen. 6.154 (= Sanzi 2003, 90: Isis no. 11.3); 
Minucius Felix, Octau. 22 A; Tertullian, adv. Marc. 1.13:. .. Osirin, quod semper sepelitur et 
in vivido quaeritur et cum gaudio invenitur; Lactantius, inst. div. 1.21.20; Firmicus Maternus, 
de errore 2.2f; etc. 

379 IGX.2, 108 11. 3f = Totti no. 72 = RICIS 113/0506: kou i£x>xeiq epa-cnv ? Iotv 
ev ctyXaian;; cf. R. Merkelbach, Zwei Texte aus dem Sarapeum zu Thessalonike, %PE 
10 (1973) 45-54 at 45-9. 



302 CHAPTER FOUR 

which was dressed and treated as Osiris. 380 This done, there followed, 
at least at Rome, three further days of rejoicing at Isis' success, culmi- 
nating in the Hilaria on 3 November. 381 

In this connection, Firmicus Maternus cites writers who gave the 
myth a rationalising twist by linking the search directiy with the natu- 
ral cycle: hanc volunt esse mortem Osyridis cumfruges reddunt, inventionem vero 
cum fruges genitalis terrae fomento conceptae annua rursus coeperint procreatione 
generari. 382 This scheme is clearly bound up with the cycle of Egyptian 
agriculture: before the building of the modern dams, the Nile began 
to rise around the summer solstice in June and reached its maximum 
height at the end of August/early September. 383 By the middle to end 
of October the river had returned to its bed; the redistribution of land 
and the sowing took place as it subsided. 384 Given the heat of Egypt 
(the new growing season for olives and vines began in January), this 
meant that the first shoots would have appeared about the time of the 
Hathyr festival. 383 



380 Delside 39, 366f., cf. Dunand 1973, 1: 236f. However, the detail of going to the 
sea implies either an immediate source set at Alexandria, Canopus, or elsewhere along 
the coast of the Delta ('il est probable que le culte de la deesse protectrice des marins 
etait pratique sur tout le littoral', Dunand 1973, 1: 114, though there is no reason to 
follow her in thinking that there was also a cult of Isis on Cape Lochias, cf. Malaise 
2005, 149f), or we might envisage a source familiar with one of the rituals observed 
in Greece or the coast of Asia Minor, say at Eretria, Cios in Bithynia or Byzantium. 

381 On the 2 Nov. the ter novena was held, perhaps sung by three choirs of nine 
members: Wissowa 1912, 354. 

382 "They claim that the death of Osiris coincides with the time of year when the 
seed is placed in the gound, and the 'discovery' of the god with that when the grain 
that has germinated in the soil's gentle warmth begins to form, in keeping with the 
annual cycle of growth": De errore 2.6., retaining with Turcan the reddunt of the ms. (see 
the apparatus on p. 80); cf. Athenagoras, leg. 22. This theory bears a family resemblance 
to those offered by Plutarch, De hide 32, 363de; 33, 364ab; 38, 366ab with Griffiths 
1970, 419f; 424; 445; see more fully Turcan 1982a, 184f. As god of the dead, Osiris 
had the power to renew life, and gradually acquired the attributes of both H'apy, the 
Nile-god, and Neper (or Nepri), the corn-god. 

383 See the lucubrations of Aelius Aristides, Or. 36.19—63, which gives an excellent 
idea of ancient speculations in this area. One of the major Egyptian festivals of Isis 
and Osiris was the feast marking the beginning of the flood proper on 1 Thoth =19 
July (Heliodorus, Aethiop. 9.9); so far as we know, however, it was not celebrated outside 
Egypt: Dunand 1973, 1: 217-19. 

384 N. Lewis, Life in Egypt under Roman Rule (Oxford 1983) 115f; PW. Pestman, The 
New Papyrological Primer'' (Leyden 1 990) chart on p. 31 4; S.J. Seidlmayer, s.v. Nil, DNP 8 
(2000) 942-4; Meeks and Favret-Meeks 1996, 168f In favoured areas with the means 
of reserving water, two crops could be sown each season. 

383 Preparations for the grain harvest in Egypt began in February; the grain was cut 
in late March-April. In Egypt there were traditionally only three seasons, each related 
to the level of the Nile: Inundation; Sinking; Deficiency (Bickerman 1980, 40). 



THE RITUAL SYSTEMS 303 

There is however good reason to think that different views of the 
significance of the festival co-existed, both within Egypt (for example, 
there was another 'finding' at Heliopolis, another at Soknopaiou Nesos, 
above Lake Moeris in the Fayyum) 386 and without, where such diver- 
gence was prompted, at least partly, by the disparity in climatic condi- 
tions between (Upper) Egypt and those of the middle-Mediterranean. 
Plutarch reveals that an interpretation such as Firmicus Maternus 
records was counter-intuitive at the latitude of Greece and Italy, where 
of course November meant the onset of colder, cloudy weather: he offers 
an account that associates the disappearance of Osiris with a series of 
analogous 'disappearances': the fall of the Nile, the cessation of the 
Etesian winds, the lengthening of the nights after the equinox, and the 
trees' loss of leaves. 38 ' None of the last three items applies with any 
force to the valley of the Nile. There were other views too: Pausanias 
was told by a Phoenician that the search for Osiris and Isis' lamenta- 
tion took place as the Nile flooded, and that 'many of the indigenous 
Egyptians' say that the flood is caused by her tears; Minucius Felix and 
Lactantius thought she mourned for her baby son. 388 

If the meaning of the festivals varied from place to place, so did the 
manner of celebration. In parts of the Aegean, the procession centred 
upon the carrying of torches, and was called (at least in Priene) r\ 
XcxiajtcxSem. 389 From Pausanias we learn that some 110 stades (= 66,000 
Greek 'feet') from Tithorea in Phthiotis, on the northern face of Par- 
nassus, there was an especially holy (ayicoxaxov) sanctuary of Isis that 
could only be visited by special invitation through a dream sent by the 



386 That at Heliopolis is mentioned by Plutarch, de hide 52, 372c, in the context 
of the solar associations of Osiris; the Charmosyna at Soknopaiou Nesos: BGU 1 no. 1 
1. 23 with Bilabel 1929, 14; 34f. 

387 Plutarch, De hide 33, 366ef, with Griffiths 1970, 448. Although various winds 
were so called (Pliny, HN2.121, cf. A. Rehm, s.v. Etesiai, RE 6 [1907] 713f.), the best- 
known Etesian winds are an Aegean phenomenon. They were generally reckoned to 
begin around the heliacal rising of Sirius (= Sothis) in mid-July and last 30—40 days. It 
is possible that Plutarch's source was thinking of the various theories, beginning with 
Thales', that the Etesians caused the Nile flood by pushing back the waters from the 
Delta, cf. Rehm 1936, 579—81. This again, however, is an external perception. The 
Tropic of Cancer crosses Egypt around 120 km south of Aswan; 25° N of latitude 
passes through Edfu. 

388 Pausanias 10.32.18; Min. Felix, Octav. 22.1, cf. Lactantius, Inst. div. 1.21.20. 

389 Athens: IGll/lW 4771 = RICIS 101/0221 (c. AD 120); 4773 = RICIS 101/0226; 
Delos, Serapeum C: RICIS 202/0209 col. i 11. 7, 22, 27; Priene: IPriene no. 195 1. 13, 
14 = RICIS 304/0802. 



304 CHAPTER FOUR 

goddess (10.32.9). 390 There were however two grand three-day festivals, 
one in spring, the other in autumn (perhaps roughly corresponding to 
the navigium and the Isia, but different in duration): on the first day, the 
sanctuary was cleansed in a manner that may not be revealed (xpo7iov 
Tiva &7t6ppr|xov) and cleared of all animal offerings remaining from the 
previous festival, which had to be buried at a distance. On the second 
day, the stall-holders build their stalls; in the morning of the third day 
the fair is held (including a cattle and a slave market); the afternoon 
is given over to sacrificing, at which the offering of sheep, goats and 
pigs was forbidden. 391 The rich sacrificed cattle and deer, the poor 
less expensive victims, koc1 %fjvoc<; kou opviGcxq xhc, ueXeaypiSaq, actu- 
ally geese and guinea-fowl. 392 Although the text at the crucial point is 
lacunate, some victims at any rate were wrapped in decorated strips of 
linen or byssos (Aivoi) TeA,a|icooiv r\ fivaaov); some were holocausted. 393 
These practices seem to reveal a fairly complex fusion of Egyptian and 
Greek elements: the manner of purifying the otSmov seems to imply a 
debt to the cult of Demeter, whereas the detail about the restrictions 
on the types of animals to be killed and the allusion to linen bandages 
suggests some attempt to adopt Egyptian custom. 394 We should assume 
that in other places too there were analogous departures from what we 
might suppose to be the 'norm' for the Egyptian cults. Another type of 
integration between Egyptian and Greek institutions is represented by 
the temples granted the privilege of aavXia, such as that of Serapis 
and Isis at Mopsuestia in eastern Cilicia, that was (to) oeu7ivoxa]Tov 
ml ev8o^ov ev 7tdar|i x[fji noXzi] kou %copou So^oc^oiievov, 395 or the 



590 This sanctuary was evidently quite different from the temple of Serapis, Isis and 
Anubis in Tithorea itself, where slaves might be freed by fictitious sale: RICIS 106/0402— 
08, 041 2f. (all dated to early II P ). The temple of Isis has never been located. 

391 Those who had eaten the flesh of sheep and goat were not allowed to sacrifice 
in the shrine of the Egyptian gods at Megalopolis for two days: SEG 28 (1978) 421 — 
RICIS 102/1701. 

392 Only one type of anseriform breeds in Egypt, namely the alopochen aegyptiaca, the 
Egyptian goose (e.g. Herodotus 2.72; Aelian, MA 5.30), which was never domesticated; we 
may assume these geese in Tithorea were not Egyptian but ordinary farmyard geese. 

393 Pausanias 10.32.14—16 with the comm. of N.D. Papachatis ad loc; Dunand 
1973, 173—8; L. Bricault, Penetration et implantation des cultes isiaques en Grece 
centrale, AncWorld 32 (2001) 147-52. The date of the foundation is unknown: Dunand 
1973, 3: 41. 

394 It seems clear however that Pausanias has rather seriously misunderstood some- 
thing here. On the Egyptian rules relating to the sacrifice of cattle, cf. Herodotus 
2.41.3—6; he also records divergent rules relating to sheep and goats. 

395 SEG 44 (1994) no. 1227 1. 12f. = Rigsby 1996, 465-71 no. 217; see also the 
commentary of M.H. Sayar, P. Siewert and H. Taeuber, Asylie-Erklarungen des Sulla 



THE RITUAL SYSTEMS 305 

temple of Isis Sachypsis at Theadelphia, south-west of lake Moeris in 
the Fayyum. 396 

b. Cultic Practice 

I was anxiously excited about this fortunate outcome as I awaited the 
morning opening of the temple. Then the gleaming white curtains were 
drawn apart and we prayed to the venerable vision of the goddess (deae 
venerabilem conspectum), while a priest made the rounds of the altars arranged 
around the temple, performing the ritual with the appointed prayers and 
sprinkling water from a libation-vessel filled from within the sanctuary. 
The rites had been duly consummated and the worshippers were loudly 
announcing the first hour of the day . . . 

Apuleius, Met. 1 1, 20 tr. Hanson 

We are familiar with the idea of Lucius as a guide to the Egyptian 
cults. In relation to the daily cult of these divinities, we are lucky 
enough to dispose of a number of other sources that, taken together, 
allow us to reconstruct both them and the organisation and composi- 
tion of the associations of worshippers. The sanctuaries of these gods 
were often quite sizeable enclosures, as at Delos (Serapeum C), the 
Serapeum at Ostia or the temple of Isis at Pompeii, with spaces for 
worshippers to gather and walk about. 397 When the temple closed, life 
continued peacefully within the enclosure, since it afforded at least 
temporary accommodation in the 7iaoxo(popiov for ^dicopoi, vecoKopoi, 
or aeditui. 396 At least in larger temples, there were often other residents, 
whether temporary or long-term, who found shelter or asylum there: 



und des Lukullus fur das Isis- und Sarapis-Heiligtum von Mopsuhestia (Ostkilikien), 
T)che9 (1994) 113-30. 

396 Bernand, IFayoum 2, nos. 112f. - Rigsby 1996, 554-56 no. 221 (93 BC). There 
was another inviolable temple of Isis Eseremphis, jointly with Heracles Callinicus, at 
Theadelphia: Bernand, IFayoum 2, 114 = Rigsby 1996, 556-9 no. 222 (70 BC). 

397 Delos: Roussel 1915-16, 250ff; Dunand 1973, 2: 87; 93-5; Ostia: G. Becatti, Scam 
di Ostia, 2 (Rome 1954) 77-85; Floriani Squarciapino 1962, 21f. with fig. 2; Malaise 
1972b, 78f. no. 66; cf. Mar 1992; Pompeii: Tran tarn Tinh 1964, 30-39; Malaise 1972b, 
275-78; Egelhaaf-Gaiser 2000, 185-95 with Abb.6 (unnumbered pages at end). The 
largest such temple in the west was certainly the Iseum Campense, of course, although 
its precise dimensions, or even shape (except for the exedra, which is on the Severan 
Forma urbis), cannot be safely reconstructed: Lembke 1994, 25 (Rekonstruktionsversuch); 
255 (Fundplan). However, the largely open area enclosed by the perimeter wall was 
c. 140m long and c. 60m wide at the southern (exedra) end. 

398 Cf. ID 2124 = RICIS 202/0296 (Serapeum C, Delos, c. 112/1 BC); RICIS 
202/0424, Face A col. if 19: ev Jtacyroipopiq) . . . (inventory, 156/7 BC). Malaise 1972a, 
141 rightly doubts whether priests actually lived in most Greek and Roman Isiac temples, 
as they traditionally did in Egypt itself (albeit on a sort of shift basis). 



306 CHAPTER FOUR 

the Katoxoi, 399 eyKaxrioavxef;, 400 iep6Soi)A,oi, 401 children dedicated to 
the gods, for example at Hyampolis and Tithorea in Phthiotis, 402 those 
awaiting initiation, those vowed to periods of chastity. 403 

Apuleius' detailed descriptions allow us an insight into virtually all 
the activities that went on day by day in a temple. 404 It is therefore a 
fairly straightforward matter to describe them. 405 



! " kcxtoxoi of the Egyptian gods are known almost exclusively from Egypt, however, 
esp. the famous case of the archive of Ptolemaeus in the Serapeum at Memphis (1 72 — 52 
BC): U. Wilcken, Zu den koctoxoi des Sarapeums, APF 6 (1913) 184-212; L. Delekat, 
Kotoche. Hierodulie und Adoptionsfreilassung. Miinchener Beitrage zur Papyrusforschung 47 
(Munich 1964); D.J. Thompson, Memphis under the Ptolemies (Princeton 1988) 2 12—65. 
Just one case relating to the cult of Sarapis is known epigraphically from elsewhere, 
at Priene c. 200 BC: IPriene no. 195 1. 29 = RICIS 304/0802: [x]oT<; KaxExouivoi.; wco 
xoC Geot), cf. Debord 1982, 92-4 (arguing that the institution here may more nearly 
resemble cases in temple estates in Asia Minor and Phoenicia than the model of 
Memphis. But this is pure speculation. At any rate, the Egyptian usage seems never 
to have been adopted in the west). 

400 Known only from IGR IV 1403 = ISmyrna no. 725 = RICIS 304/0204 (Smyrna, 
AD 211), a dedication by the philosopher Papinius, eyKaTOxriaai; x& Kvpicp LapcauSt. 
It may simply be a variant of kotoxoi;, though the high status of the man involved (he 
was able to obtain permission for extending the Nemeseion directly from Caracalla) 
perhaps makes this unlikely. 

401 IG XIV 1024 = IGUR 101 = RICIS 501/0107 (Rome, early IIP): iepoSoutax; 
(1%) JtdoiK i£po8ot)Xi<a><;; cf. IG XIV 914 = Malaise 1972b, 90f. Portus Ostiae 6 = 
RICIS 503/1211 1. 20f 

402 Hyampolis: IGlX.l no. 92 = RICIS 106/0302 (II-P); Tithorea: IG1X.I no. 187 
= RICIS 106/0401 (early IP). 

403 Cf. Tibullus 1.3.22-32; Propertius 2.33a.lf; 4.5.34; Ovid, amor. 2. 13. 17f; Juvenal, 
Sat. 6.535—41; Apuleius, Met. 11.19, castimoniorum abstinentiam satis arduam; also Dunand 
1973, 3: 192; Merkelbach 1995, 137f; J. Alvar, Marginalidad e integration en los 
cultos mistericos, in idem and F. Gasco (eds.), Heterodoxos, reformadores y marginados en la 
Antiguedad Cldsica (Seville 1991) 71-90. 

mi "p^g re l a ti on between the daily cult and the well-known frescoes from Hercula- 
neum (see Pis. 23 and 24 here) is quite uncertain. Moreover, their provenance is com- 
pletely unknown. I therefore allude to them for particular details only. The bibliography 
is: Tran tam Tinh 1971, 83f no. 58 fig. 40 and pp. 29-38 = Malaise 1972b, 251f. no. 3 
with pi. 35 = Merkelbach 1995, 553 Abb. 553 and colour-plate IV (Isiac ceremony with 
choirs); 85f no. 59 fig. 41 with pp. 39-41 = Malaise 1972b, 252f. no. 4 = Merkelbach 
1995, 554 Abb.73 and colour-plate V (between pp. 324-5) (Isiac dance). 

405 It is unclear what relation there was between the daily ritual in the Graeco-Roman 
world (assuming it was fairly standardised) and that of temples in Egypt, where the 
'Opening of the Mouth' and/or the abridged 'Opening of the Mouth for breathing' 
were of central importance (Smith 1993, 13—18). Malaise 1986a, 89ff argues that the 
Egyptian liturgy was used, but in my view the rituals seem similar to, though perhaps 
somewhat more elaborate than, those that were performed in other Graeco-Roman 
temples (cf. Scheer 2000, 65). Temples in Egypt held three different services, morn- 
ing, noon and evening (of which the first was by far the most important), whereas 
Isiac temples in the Mediterranean world noted only the moments of opening and 
closing: J. Vandier, La religion egyptienne. Mana 1 (Paris 1949) 164—7; Sauneron 1957, 



THE RITUAL SYSTEMS 307 

The text at the beginning of this section mentions the first event of 
each day, the opening of the temple. A sizeable group of worshippers 
was already gathered in front of the doors, just as there was when the 
doors closed again in the evening. The service began when the god 
'awoke', a solemn fiction supervised by the cbpoJioyoq, whose duty it was 
to announce the appointed time for the performance of every ritual 
act. 406 The next step was for the statue to be washed and dressed. This 
demanded an entire staff, in particular the (male) oxokioxai, and the 
female ornatrices.* 01 They bathed the statue and combed its hair, 408 per- 
fumed it, burned incense and poured libations before it. The axoXiaxai 
then arrayed the statue in costly robes and hung it with jewels given by the 
better-off worshippers, to register their devotion and their circumstances. 
We have a number of striking examples of such generosity, including 
objects in gold (crowns, tiaras, diadems, fasteners, rings, ear-rings) as 
well as precious stones, not only from the well-known Delian inventory 
lists but also through votive-inscriptions from Pergamum, Thessalonica, 



75-86; Dunand 1973, 1: 189-207; Meeks and Favret-Meeks 1996, 126-9. Note also 
the description of the daily ritual at Edfu by Alliot 1949, 1-197. 

""' A (bpoXoyiov (sundial, clepsydra, or some other device for measuring time) is 
mentioned as a dedication in ID 2087 = RICIS 202/0342 1. 1 1 (Serapeum C, Delos) 
and AE 1972: 168 = RICIS 509/0201 (Helvia Ricina, Picenum); this latter is inscribed 
on a spherical sundial. At least two Egyptian klepsydrai dating from the time of Alex- 
ander the Great are known, one in the British Museum, the other in the Hermitage, 
inv.2507 a: Ausstellung Liebighaus, 548f. nos. 112f; cf. Lembke 1994, 248, cat. no.**E55. 
Two early-Ptolemaic examples belonged to the furniture of the Iseum Campense: 
Lembke 1994, 246f. cat. nos. E52f. (one now in Turin, the other in the Museo Bar- 
racco). Another such water-clock, which were needed to tell the time at night, has 
been discovered at Ephesus: G. Langmann, G. Holbl and M. Firneis, Die agyptische 
Wasserlaufuhr aus Ephesos, JdOI 55 (1984) Beibl. 4-68. Clement, Stromat. 6.4.36 men- 
tions that the cbpoXoyoi; processed holding a (bpoXoyiov. On the importance of time 
in the Egyptian cults, cf. Malaise 1978, 685f. According to Porphyry, de abst. 4.9, the 
i)|j,vo)OT|<; woke the statue by speaking to it in Egyptian after he had purified the cella 
with water and lighted the sacred fire (explicitly with reference to the cult of Serapis, 
either at Alexandria or Memphis, cf. Patillon-Segonds 3: 61f; Dunand 1973, 3: 198; 
Merkelbach 1995, 150f. §276). 

407 aioXiaxai: e.g. Plutarch, de hide 3, 352b4; 39, 366f2; IG II/IIP 3644 = RICIS 
101/0215 (Athens), IG XIV 2338 - Pais, CIL V Suppl. 226 = RICIS 515/0125 
(Aquileia); ornatrix f (am?): CIL XII 3061 = RICIS 605/0103 (Nemausus). 

1(18 IG V2, 472 = RICIS 102/1702 (Megalopolis, II-IIP), with F. Dunand, Sur une 
inscription isiaque de Megalepolis, Z PE l (1967) 219-24; eadem 1973, 2: 164f.; cf. 
Apuleius, Met. 11.9: et quae pectines eburnos ferentes gestu bracchiorum flexuque digitorum ornatum 
atque oppexum crinium regalium fingerent (women in the Isiac procession, evidently referring 
to the daily ritual in the temple). 



308 CHAPTER FOUR 

Tarracina, Nemi, Italica in Baetica, and Acci (Guadix) in Tarraconensis. 409 
Once the statue had been washed and adorned, it was offered food that 
had been brought by the worshippers who had risen early to hand in 
their offerings as soon as the temple-doors opened, and then cooked in 
the temple. These dishes were later actually consumed by the priests 
and servants, as in Greek and Roman temples. 410 

The temple-area remained open and accessible to worshippers all 
day long. Given the ritual demands made on individuals, there must 
have been a good deal of coming and going. The doors were closed 
again in the afternoon, and a ceremonial performed similar to that of 
the morning, albeit less elaborate (Sauneron 1957, 87). Martial recounts 
how the priests formed up in a group to announce to the goddess that 
the eighth hour had arrived, which meant that the temple was clos- 
ing and that the worshippers present had to take their leave. 411 There 
must also have been a certain amount of commercial activity in the 
outbuildings, since they often housed stalls for flower- and fruit-vendors, 
workshops for sculptors, coroplasts, lamp- and amulet-makers, and all 
the workers who were needed to keep the temple-enclosure and its 



409 Delos: RICIS 202/0423, 0427, 0433; rings in other lists: 202/0431; 0428 1. 29; 
Pergamum: IPergamon 336 = RICIS 301/1202 (three white linen robes, 80 gold leaves, 
cf. Dunand 1973, 3: 93-6); Thessalonica: IG X.2, 114 = RICIS 113/0556 (jewelled 
ear-rings, II 1 '); Tarracina: CIL X 6303bis = RICIS 502/0701: cum collari argenteo; Nemi: 
CII XIV 2215 = IIS 4423 = RICIS 503/0301: ..basikum ornatum ex gemmis n(umero) 
I. . . collarem ex gemmis beryllis, spatalia cum gemmis II, collarem alterum cum gemmis n(umero) 
VII, inaures ex gemmis n(umero) X. . . vestem liniam: tunicam I, pallium I, zona I cum segmentis 
argenteis, stola I. . . (P or earlier); Italica: AE 1 982: 521= RICIS 602/020 1 : cum inauribus 
tri[bacis marjgaritiis n. X et gemmis n. XXXX et berull(is) n. VIII et corona aurea cum gem(m)is 
n. XXV et gem(m)areis (VII?)...; Acci: CIL II 3386 = ILS 4422 = RICIS 603/0101 (gift 
by Fabia Fabiana, second half II P , too long to cite here). On the Spanish texts, see 
J. Alvar, La sociedad y el culto: Isis en la Betica, in C. Gonzalez Roman (ed.), La Sociedad 
de la Betica. Contribuciones para su estudio (Granada 1994) 9-28, with the warning that CIL 
II 3387 is probably a modern imitation of the Fabia Fabiana text. 

410 According to the sacred law from Priene (IPnene no. 95 1. 29 = RICIS 304/0802 
1. 26-29; see n. 399 above), the food-offerings deposited at the temple by the Sf]|j,o<; 
are to be fed to the Kax£x6|ievoi two zox> 9eoi>. For the use of the food deposited as 
TpOOT£^(0|j,aTa or mensae in Greek and Roman usage, see B. Gladigow, Zur Ikonographie 
und pragmatik romischer Kultbilder, in H. Keller and N. Staubach (eds.), Iconologia 
sacra: Mythos, Bildkunst und Dichtung in der Religions- und Sozialgeschichte Alteuropas (Berlin 
1994) 9-24; Riipke 2007a, 141. 

411 10.48.1. It is difficult to give an equivalent time in our reckoning, since the length 
of hours varied with the season and the latitude (Bickerman 1980, 15); the ninth hour, 
at any rate, was dinner-time (Martial 4.8.6), and richer people generally stopped 'work' 
at the seventh hour. This implies that the temples actually closed not at dusk but in 
the early-mid afternoon. 



THE RITUAL SYSTEMS 309 

various buildings in good repair and to supply worshippers with what 
they required went they went into the sacred area. 412 

There were however restrictions upon entry. With regard to Egypt 
itself, we know something of such rules from a ceremonial hieroglyphic 
text, actually inscribed in the time of Hadrian or Antoninus Pius 
on the columns of the hypostyle of the temple of Chnoum at Esna 
(Latopolis), 54 km south of Luxor, on the basis of older texts in the 
temple library 413 The rules are included among the regulations for the 
festival of the advent of the creator-goddess Neit(h) on 19-20 Epeiph 
(mid-July Nile-flood), but can perhaps be taken as generally valid, if 
not in detail, then in principle, for Egyptian cults generally 414 There 
are three sections. The first applies only to persons wishing to take 
part in processions. 

All the men must have abstained from sexual contact with a woman 
for one day. They must be purified; they must be washed; they must be 
(appropriately) dressed. 

The second section applies to those who are admitted to the temple 
courtyard: 

No one may be admitted to the temple who is possessed or bewitched: 
their place is on the perimeter of the sacred area. . . . 
Those who remain outside the temple are to remain seated to the right 
and left (of the main avenue), but are not to go to sleep. People are per- 
mitted to exclaim with joy anywhere in the outer temple. 

The third and longest section applies to those permitted access to the 
temple-interior: 

The offering shall be placed on the festal altar of this holy divinity by the 
prophets, the purifier-priest and all the usual personnel of the temple. 
No one in mourning shall be admitted into the temple cella. 
All those who enter must have removed his body-hair, cut his nails, be ton- 
sured. Anyone who wishes to enter the cella must be dressed in fine linen; 
anyone who wishes to enter the cella must be purified with natron. 



112 Cf. Juvenal, sat. 12.27f.: pictores quis nescit ab hide pasci? 

415 Sauneron 1962, 1—5. The texts include a calendar of festivals, and a series of 
ritual instructions. Both are probably, at least in basic form, a good deal older than the 
date at which they were inscribed: Sauneron speaks of the New Kingdom. 

111 It is frankly unknown how far such rules were observed outside Egypt. Briefly 
on Neit, see Bonnet 1952, s.v. Neith. 



310 CHAPTER FOUR 

Anyone who wishes to enter the cella must have avoided sexual contact 

with a woman for eight days; he must have abstained from impure foods 

for four days. 

Every man who wishes to enter this temple, or who needs to accomplish 

some ritual there, must depilate his head and body and cut his nails, 

and worship the god in the main aisle, in the area reserved for the 

townsfolk. 

As regards the normal temple-personnel, who are above these (i.e. allowed 

to approach nearer the altar?), they must keep themselves from contact 

with a woman for nine days, and pure from everything else for five. 

Those who fulfil these conditions are permitted to gain access to the 

temple via the side gate of the pylon, after they have washed in the lake 

and donned their clothes once more. 

All Asiatics, young and old, are prohibited from entering the temple! 

Women are prohibited from entering any part of the temple-area closer 

than 200 setat on any side; 41 ' it is a sin to approach nearer than 1500 

cubits from north or south, or 1500 cubits from east or west. 416 

There are further restrictions on the consumption of wine, though they 
give the impression of being short-term, related to the forthcoming 
encounter with the god. Lucius too avoided wine during the prepara- 
tions for his first initiation (Apuleius, Met. 11.23). Otherwise, however, 
wine was of course used in libations; the sacred law from Priene speci- 
fies that the excess wine from the ceremonies was to be a perquisite of 
the priest of Serapis. 417 Analogous rules to those at Esna, for example 
on sexual purity and abstinence from certain foods, can be guessed 
to have been imposed by Egyptian temples in the Mediterranean world, 
particularly upon priests. An unfortunately extremely fragmentary sacred 
law from Iasos on Rhodes begins: dtp' cbv Set ayv[6]v eie>7top£'u£a9oct. . . , 
"Those who enter must be free from the following impurities...". 418 



*'" Two notices in Serapeion A at Delos from shortly after 166 BC likewise prohibit 
the entry of women (and of men dressed in anything made of wool): ID 2180f. = 
RICIS 202/0199-200. 

416 Sauneron 1962, 340—9 (with the explanatory notes); cf. L. Kakosy, Probleme der 
Religion im romerzeitiichen Agypten, ANRWIIA&.5 (1995) 2894-3049 at 3023f. 

117 IPriene no. 95 1. 29 = RICIS 304/0802 1. 26 (see nn. 399, 410 above). Of course 
this is not necessarily to say he was free to drink it up; such emoluments were normally 
to be sold on the market. As Malaise 1986a, 81 observes, restrictions on wine-drink- 
ing were part of the separation from the world thought to be necessary before close 
contact with the divine. 

418 I.Iasos 2, no. 242 = RICIS 305/1403 1. 1 (Principate); lines 2-14 are illegible, 
and the statements by Dunand 1973, 3: 192f. about their content, based on Accame's 
reading, cannot be sustained. The ruling about entering the temple in SEC 22 (1967) 
114 = LSCG no. 50A = RICIS 101/0401 (Teithras, Attica, late I") seems to refer to the 



THE RITUAL SYSTEMS 311 

Stories were circulated about the consequences of disregarding such 
rules, and thus tended to reinforce them. 419 

Rules of this type tended to emphasise still further the alterity of 
the temple, already clear in its architectural and decorative differences 
from the Graeco-Roman norm, and in some cases, most especially the 
Iseum Campense, enhanced by Egyptianising decoration: lotus-leaf 
capitals, figural, unfluted granite columns, puzzling statuary, obelisks 
and hieroglyphs. 420 Equally strange was the appearance of the clergy, 
dressed as they were in white linen robes, their heads and cheeks shaved: 
grege linigero.. . et calvo (PI. 24). 421 The rattling of the sistrum, taken to be 
a specifically Isiac instrument, was notorious. One of the Herculaneum 
frescoes shows a masked Ethiopian performing an apparently wild 
dance on the temple steps before admiring spectators, all dressed in 
exotic robes (PL 23). It was easy enough for sectors of Roman public 
opinion to equate this alterity with corruptibility in order to discredit 



party involved in the dispute that the text is attempting to resolve, and can probably 
not be generalised. It is somewhat troubling that sacred laws similar to that from Iasos 
have not been found; one may however reflect that what gets noted in such texts is 
the new, and it may well be that such rules for worshippers were so widespread, and 
so enforced by the ^aKopot, that they did not need to be inscribed. 

419 E.g. Pausanias 10.32.17, cited n. 45 above. 

1211 Best appreciated in the plates to Lembke 1994, illustrating the unfortunately 
very limited remains of the decoration of the Iseum Campense (which was however 
extreme, indeed, with the partial exception of Beneventum, unique, among Isis- and 
Serapis- temples in the Greek and Roman world in its search for 'authenticity'). She 
aptly notes that none of the items actually imported from Egypt were chosen because 
they had anything to do with Isis or Serapis (they do not), they were simply decoration 
("einen musealen Charakter verliehen haben muB"), and taken mainly from the Delta 
area, handy for transport (pp. 33—50). The same is true of Beneventum. However the 
temples at Herculaneum, Pompeii and Ostia all made some effort to emphasise their 
alterity; for example, one of the pylons at the entrance to the Serapeum at Ostia was 
decorated with an Apis-bull in pumice: Floriani Squarciapino 1962, 21 with pi. VII; 
cf. more generally Wild 1984; Bommas 2005; Quack 2005. The same point emerges 
clearly from the Herculaneum frescoes: in the 'Procession' fresco (PI. 24 here), sphinxes 
are placed on either side of the temple-entrance, palm-trees grow, ibises wander about; 
in the 'Dancer' fresco, as well as the palms and the temple-decoration, there is an 
'oriental'-looking altar in front of the steps (PI. 23). The vogue for Egyptianising deco- 
ration in private houses no doubt familiarised these images to some extent, however; 
cf. J. Leclant, Aegyptiaca et milieux isiaques. Recherches sur la diffusion du materiel 
et des idees egyptiennes, ANEW II. 17.3 (1984) 1692-1709; Merkelbach 1995, 134-7, 
with the corresponding plates; C. Maderna, Agypten — phantastische 'romische' Welt, 
in Ausstellung Liebighaus, 434-45. 

121 Juvenal, Sat. 6. 533; cf. Martial 12.28.19-21, Minucius Felix, Octav. 22.1; Arnobius, 
adv. nat. 7.33 etc. On the perceived alterity of Egyptian priests and cult, cf. Malaise 
1972a, 385-406. 



312 CHAPTER FOUR 

the Egyptian cults, most famously, as we saw in Chapter 3.3, in relation 
to the exploitation of the supposed credulity of women. 422 If anything, 
however, it was not the priests or the sistrum but the theriomorphism 
of some Egyptian divinities, above all Pharia iuvenca and latrator Anubis, 
that caused the most enduringly hostile comment. 423 

It was however not so much the daily ceremonies of opening and 
closing the temple as the manifold activities undertaken in it or near it 
that helped to bind worshippers of the Egyptian gods to one another 
as a group and, as individuals, to construct appropriate sentiments and 
dispositions as regards these gods. One such event is the sacred dancing 
I have just mentioned. Another is likely to be smaller processions staged 
on occasions other than the Isidis navigium. A probable example is the 
second of the Herculaneum frescoes (PI. 24), which depicts the begin- 
ning of a procession no doubt analogous to the grander Alexandrian 
one described by Clement [Strom. 6.4.35), albeit on a far smaller scale, 
with the 7tpo(pr|Tr|(;, the head of the temple-organisation, emerging from 
the temple carefully carrying a hydria, the ultimate reference-point of 
the Egyptian cults abroad, and greeted by two antithetical choirs, before 
beginning the sacred tour (cf. Malaise 1972b, 25 If). The deliberate 
alterity of such events exercised an undeniable fascination on many of 
the spectators, a fascination conveyed both by the 'documentary' detail 
the painter has chosen to include and by the focusing of the spatial 
organisation upon the figure of the Jtpo(pr|Tr|c; himself, and the choice 
of gold-paint to pick out the hydria. 

I would like at this point to examine the role of two basic ritual 
practices, sacrifice and prayer, as elements in the process of binding 
worshippers to the Egyptian cults. 



422 Josephus,4J 18.3.4, §65-80; cf. Juvenal, sat. 6.488f.; 525-41 etc. Of the story of 
Deems Mundus and Paulina (the affair of AD 19), Malaise comments, "Sans doute, 
dans ce recit . . . le seul detail veridique est-il le debut d'une participation ouverte de 
matrones au culte isiaque": 1972a, 88; cf. Heyob 1975, 118f; Mora 1990, 2: 91-5 
seems by contrast to accept the story as veridical. 

423 Cicero, de nat. dear. 3.47; Vergil, Aen. 8.698; Propertius 3.11 .40-6; 3.28a. 1 7f.; Ovid, 
ars amat. 3.635; Lucan, bellciv. 8.831, Plutarch, de hide 71, 379de etc.; cf. the enormous 
compilation of passages by Smelik and Hemelrijk 1984. Leclant 1981, 873 rather oddly 
plays this aspect right down; Richter 2001, 208f. rightly refers to the general argument 
of Plutarch, de superst. as consistent with this rejection. Apuleius, Met. 11.11 claims that 
the aim of strangeness in the Egyptian cults was etiam ipsa novitate reverendam. 



THE RITUAL SYSTEMS 313 

i. Sacrifice and Votives m 

The most prominent victim in the literary sources is the goose, 425 which 
also seems to have been in some sense a sacred bird — at any rate there 
was a flock of them in the grounds of the temple of Isis and Serapis 
at Smyrna. 426 Here too however there is clear evidence of adaptation 
to and integration with Greek and Roman sacrificial customs. Thus 
the lex sacra from Priene prescribes the offering of two chickens by the 
priest at the Isia on 20 Apatourion, as well as a number of other things: 
a 'quarter' (xexapxeit;) of barley two of wheat, seven bronze spits; the 
sanctuary is also to be purified with the blood of a chicken. 427 The 
altar of the temple-caretaker (aeditimus) Astragalus now in the Louvre 
shows him about to sacrifice a dove or possibly a partridge to Isis over a 
lighted altar with fruits piled on it. 428 A fragmentary Hadrianic sacrificial 
calendar from Athens now in the Ashmolean Museum, probably that 
of a private religious association, lists a sacrifice on 13 Boedromion 
(mid-September) to the couple Nephthys and Osiris of a cockerel and 
a selection of fruits; at the same ceremony, barley and wheat grains 



424 For a list of ritual objects in the Egyptian cults, including fumigations, means of 
illumination and musical instruments, see Dunand 1973, 3: 218—21. 

425 Ovid, Fasti 1.453f [Inachiotis = Isis); AnthPal 6.231 1. 4 — Garland, Philip no. xxc; 
Juvenal, sat. 6.540f; Aelius Aristeides, Or. 49 (= Sacred Tales 3) 45 (the seller of geese is 
instructed by Isis to reserve his last two birds for Aristeides); a goose is depicted beside 
the Isiac priest in the illustration for mensis November in the calendar of Filocalus: Stern 
1968, 50 fig. 5; cf. Dunand 1973, 3: 76; 205. Documentary evidence in Egypt: a goose 
among the gifts to be given to the King eii; to Eiateia, i.e. at the Hathyr festival of 254 
BC: Apollonios to Zenon: P.Cair.^en. 4, 59560 1. 4fi; declaration of Horus about the 
delivery of a goose and goose eggs (?) eii; to Eicufja: BGU VII 1501 (Philadelphia in 
the Fayyiim); for Tithorea, see n. 392 above. The Egyptian goose, which was not much 
eaten (indeed in modern times has been considered uneatable), seems simply to have 
been one of the commonest sacrificial victims in Egypt: Bonnet 1952, 200, s.v. Gans; 
contra: L. Stork, s.v. Nilgans, LAg 4 (1982) 484. Harpocrates is sometimes represented 
as riding on, or simply holding, a goose (also cockerels, ducks, dogs, sphinxes etc.), 
e.g. Dunand 1979, 232f nos. 222-26; Bayer-Niemeier 1988, 104 no. 133 = Ausstellung 
Liebighaus, 654 no. 243; Bayer-Niemeier 1988, 69 no. 29 = Ausstellung Liebighaus, 659f. 
no. 253; Torok 1995, 76f. nos. 84f (identified as a duck, however, which I believe to 
be mistaken, esp. as the type was surely taken over from the goose of Aphrodite). But, 
as Bonnet points out, the reason for the association is obscure. 

12fl Aelius Aristeides, Or. 49 (= Sacred Tales 3) 49 with Festugiere 1986, ad loc. 

427 IPnene no. 195 1. 9f. = RICIS 304/0802 (c. 200 BC): &uoei 6e 6 lepeix; tw[i 

lapd [7u8i . . . t5>v vopi^oiaivcov voc[oSv ouo. . .; 1. 36: [to iep]6v KaGaipeico voo[owi 

See also Dunand 1973, 3: 56f; 149f; 204; 206. 

428 OIL VI 345 = Malaise 1972b, 113f. Rome no. 6 = Lembke 1994, 141 B**6 and 
246 **E51 with pi. 47.3 = RICIS 501/0122. All are agreed that the bird is a dove; to 
me the beak looks more like that of a partridge — or even a guinea-fowl. At any rate, 
it is far too small for a goose. 



314 CHAPTER FOUR 

are to be scattered, and a iieAiicpcxxov, milk or water sweetened with 
honey, offered as a libation. 429 Such a mixture of animal and vegetable 
offerings is common in Greek sacred laws. 430 

As a class, lustrations and libations were of great symbolic importance 
in the cult of the Egyptian divinities. There is plenty of evidence that 
such Nile-water was considered to be more than merely a purificatory 
agent: it was an especially effective mediator, having the property of 
turning offerings into material assimilable by the gods. 431 Genuine or 
merely suppositious Nile water was thus freely used in ritual: not only 
are t>8peux, the vessels used to contain it, mentioned in several votives, 432 
there is archaeological evidence for such objects, 433 and situlae are 



429 IG II/IIP 1367 - LSCG no. 52 = RICIS 101/0225 11. 4-6 (extract only). This 
is the sole western epigraphic evidence for the worship of Nephthys, the sister of Isis 
and Osiris, wife of Seth (E. Graefe, s.v. Nephthys, LAg 4 [1982] 457-60), though it is 
generally thought that the deity Netoxepa conceals her (Malaise 1972a, 215f.). Since she 
is closely associated with Isis' burial of and mourning for Osiris, and was one of the 
four divine guardians of the canopic jars (see chap. 2.1.a.i above), Bricault is probably 
correct to see here a caique on Persephone and Hades, particularly given the agrar- 
ian-Dionysian associations of the remainder of the text (cf. Nilsson, GGR 2: 331), and 
of the (xeXiKpatov, cf. Graf 1980, though such a libation is also found in connection 
with holocaust-sacrifice (e.g. SylP 1035 1. 34 — LSCG 151A [Voropfer to Zeus Polias and 
Hestia on Cos]). Dunand 1973, 2: 137—9 rightly emphasises the Greek character of 
the offerings; elsewhere (3: 240f) she rather implausibly argues that this sacrifice was 
in fact part of the Hathyr festival. 

430 Round cakes (popana), along with a goose, are mentioned as an oifering to Isis 
by Philip of Thessalonike, AnthPal 6.231 1. 3 = Garland, Philip no. xxn; and to Osiris 
by Juvenal, sat. 6.541; cf. Roeder 1916, 2127. 

431 Cf. Alliot 1949, 46-9; Wild 1981; Malaise 1985; Kaplony 1989. On the very 
frequent terracottas of Harpocrates, but also Isis and 'Bubastis', with a round-bellied 
nw-pot, which is the standard Egyptian water-libation vessel, used as a figure for blessing 
and fertility, see Dunand 1979, 74f; Torok 1995, 62f discussion under no. 62. 

432 7GX.2, 83 = RICIS 113/0521 (Thessalonike, 37/6 BC (?)); ID 2617 1. 4 = RICIS 
202/0206 (Serapeum C, Delos); ID 2620 = RICIS 202/0210 1. 1 (ibid.); ID 1435 = 
RICIS 202/0425 (inventory, Delos, apparently afterl56/5 BC); hydraeum cemmis (i.e. 
gemmis) exornatum et auratum: CIL XIV 3941 = ILS 4378 = RICIS 503/0801 (Nomentum). 
Also the Isiac lacus at Lambaesis, which must have been intended as a source of such 
water: CIL VIII 2631 = 18101 = ILS 5778 = RICIS 704/0303 (mid-HP). Malaise 
1972a, 126 suggests that canopus jars fulfilled the same function in the west; e.g. the 
two diorite statues of priests carrying canopic-jars at Beneventum: Muller 1963a, 95—8 
no. 284; 106 no. 288 (this "diorite" is not true diorite but a spotty, 'interesting' stone 
["Chephren-diorite"] extracted in the Libyan desert and mainly found in worked form 
at Memphis: cf. D. Wildung, s.v. Diorit, LAg 1 (1975) 1096). 

433 Cf. Griffiths 1975, 208-10; Wild 1981, 105-10. 



THE RITUAL SYSTEMS 315 

commonly represented being carried by Isis and her priestesses. 434 
Apuleius describes an especially venerated gold urnula with a spout and 
handle being carried in procession, 433 and there is some iconographic 
evidence for similar vessels, for example on the well-known Isiac pro- 
cession in the Vatican formerly in the Maffei collection, where a sacred 
pitcher is being carried by a priest with veiled head and arms (see PI. 
22). 436 On the other hand, there is no very satisfactory evidence concern- 
ing the precise ritual significance or use of such water. Most scholars 
assume that its regenerative power was the focus of interest: the worship- 
pers gratefully offered Nile-water to the gods not only to assure agrarian 
prosperity but also the perpetuation and regeneration of human life. 437 
There are several references in Late Egyptian texts at Dendera to the 
'humours' of Osiris that were preserved in vases. 438 Libations inside the 
sanctuary may have been intended to recall Osiris' rising with the Nile- 
water: according to Plutarch, as I have already remarked, the Nile, and 



434 E.g. the well-known Isis-statue in dark marble in the Museo Archeologico in 
Naples: Tran tarn Tinh 1990, 767 no. 53a = Ausstellung Liebighaus 623 no. 205 with 
colour pi.; the equally well-known statue supposedly from the Villa of Hadrian at Tibur, 
now in the Capitoline: Helbig 4 2:238 no. 1433 = Malaise 1972b, 11 Of. Villa Adriana 
no. 9; the headless statuette of Isis in the Palazzo Altemps, inv. 4235: Collezioni egizie, 
66 no. 23 with colour pi. (otherwise unpublished); terracotta statuette of Isis with situla 
in Musee des Beaux Arts, Budapest: Torok 1995, 86 no. 104; Isis holding a nw-pot: 
Dunand 1990,144 no. 382; and the list in Tran tarn Tinh 1990, 776(e). Priestesses: Isias 
from Smyrna, grave relief now in the British Museum, inv. Sc 639 (IP): Walters 1988, 
53f — Eingartner 1991, 143, no. 98 pi. 62; Cantina Proc(u)la depicted on her funerary 
cippus (CIL VI 34726) in the Palazzo Massimo: Collezioni egizie no. 16 with colour pi.; 
the priestess from Tauromenium: Malaise 1972b, 323 Tauromenium no. 9 with pi. 
64. This item of equipment is virtually standard in the Attic reliefs showing women 
as Isis, cf. Walters 1988, passim; 2000, 64 fig. 1; 68 fig. 4, 70 fig. 5. Loose examples of 
late-period Egyptian situlae are also known, e.g. Collezioni egizie 76 no. 36. 

433 Met. 11.11; for the parallel (but partially conflicting) passages, cf. Griffiths 1970, 
437 on de hide 36, 365b2f 

436 Relief: Helbig 4 1: 388f. no. 491 = Malaise 1972b, 234f no. 441 = Merkelbach 
1995, 615 Abb.145. 

437 See Chap. 2.1. a on Osiris and the Nile. Firmicus Maternus, de errore 2.5 speaks 
of the worshippers expecting some blessing from the water each year: hanc aquam quam 
colis putas aliquando prodesse. Already in the Ptolemaic period, one of the Isiac hymns 
from the temple of Philae (in Egyptian) mentions water brought from Biggeh "that 
bestows everlasting life": Zabkar 1983, 134 (Hymn 4); Biggeh was the location of the 
most famous tomb of Osiris, the Abaton: Dunand 1973, 1: 149f. 

438 L. Pantalacci, Une conception originelle de la survie osirienne d'apres les textes 
de Basse Epoque, Gottinger Miszellen 52 (1981) 57-66 at 58f. 



316 CHAPTER FOUR 

indeed all water, is the'OatpiSoc; raioppori, just as Isis is the lord of dew. 439 
But the hydria seems itself to have become the metonymic object of 
adoration when carried in procession (e.g. PI. 24). 440 It seems even 
to have become a sort of deity Hydreios, to whom offerings might be 
made. 441 

Milk, as the liquid that enables the infant to survive after birth, may 
in some contexts have had a role analogous to water. At any rate, Apu- 
leius describes a priest at Cenchreae pouring libations of milk from a 
small, round breast-shaped vessel (in modum papillae rotundatum); later a 
mess of milk and kibbled grain was floated out onto the waters of the 
harbour. 442 The lex sacra from Priene, as well as the occasional images 
of Isis (and Serapis) with patera, shows that wine too was used for liba- 



139 Plutarch, de hide 33, 364a; 36, 365b; 38, 366a etc. with Griffiths' note on 
p. 424; Isis = dew, PGrMag XII. 234: eyco ei|j.i ^laiq f| KaXovuevri Spoao<;. For what 
it is worth, 18 amphorae inscribed lepdrcu; (gen.) Scbpa have been found at different 
locations in Pompeii: GIL IV 6546£; Tran tam Tinh 1964, 85 and 177f. nos. 150-162; 
SEG 42 (1992) 922; RICIS 504/0219. The courtyard of the Serapeum at Ostia was 
decorated with Nilotic scenes, now virtually disappeared: Floriani Squarciapino 1962, 
2 1 . Underground cisterns have been found in Serapeum A and C at Delos (cf. Dunand 
1973, 2: 101), at Gortyn and in the second temple of Isis at Pompeii: Wild 1981, 34-8; 
40-47; the tympanum of the tempietto protecting the cistern at Pompeii is decorated 
in stucco, and shows a hydria, with a kneeling worshipper to left and right (Tran tam 
Tinh 1964, 34 = Malaise 1972b, 276f; 1985, 145; Wild 1981, 76-84). Such construc- 
tions were simply means of accommodating Egyptian ideas to the different physical 
circumstances of the Mediterranean world. 

440 Vitruvius, de arch. 8 praef. 4 remarks on the adoration of the hydria in the Egyptian 
cults; Clement, Stromat. 6.4.37 describes the Jtpocprixrn; in a procession holding a hydria 
carefully against his chest, cf. Malaise 1972a, 116; a fine example of a gilt-copper 
hydria for use in Egyptian cult was found in 1831 at Egyed nr. Scarbantia (Sopron) in 
Pannonia Sup.: V Wessetzky, Die agyptischen Kulte zur Romerzeit in Ungarn. EPROER 1 
(Leyden 1961) 42f. with pis. VI-VIII, figs. 10-12 [this book must be one of the worst- 
organised academic titles published since the seventeenth century, having no list of 
contents, no indices, no sub-headings and no catalogue-numbers]. On the hydria more 
generally, see Wild 1981, 3; 69; 157f.; Malaise 1985, 144. 

441 Dedication of ritual objects by a melanephoros from Alexandria in Serapeum C 
at Delos to Serapis, Isis, Anubis, Harpocrates and 'YSpeion, 95/4 BC (?): ID 2087 - 
RICIS 202/0342; cf. Malaise 2005, 59-66. 

442 Apuleius, Met. 11.10; 16, cf. Malaise 1986, 95. The image is of course taken 
from the common representation of Isis offering her breast to the infant Horus, e.g. 
Ausstellung Liebighaus 646f. no. 233 (Allard Pierson museum, inv. 7766, from Alexandria, 
Roman period). 



THE RITUAL SYSTEMS 317 

tions, as we would expect. 443 In Egypt, the nw-pot was used both for 
water- and wine-libations alike. 444 

Another type of ritual container was the cista. In the myth, Osiris' 
head was preserved in a wicker-work casket; Isis, inasmuch as she was 
keeper of the bits of his dismembered body, was herself identified with 
it. 445 However the cista, in the familiar form of a circular wicker-work 
basket, is unknown in the Egyptian cults until the Roman period. 446 
Although they certainly did borrow other items from the Eleusinian- 
Dionysiac sphere, Malaise has argued that in this case there was none 
(1985, 135). There are only three literary allusions to cistae, and in each 
case the reference is to something different, so that genuine doubts 
have been expressed as to whether they had any specific role in the 
cult. 447 Nevertheless there is clear archaeological evidence for them, 
not merely on Isiac altars but also on frescoes at Pompeii, for example 
below the inventio Osiridis scene in the later Iseum at Pompeii. 448 The 



443 IPriene no. 195 11. 26, 32f. = RICIS 304/0802; Isis + patera: e.g. bronze statuette 
of Isis Fortuna in the Palazzo Massimo, inv. 256095: Collezioni eeizie 87, no. 53; CIL 



VI 19875 - Malaise 1972b, 128 Rome 54 = RICIS 501/0191 with pi. XC shows Isis 
with one hand raised in blessing, and holding a patera in the other; there is a common 
series of factory lamps showing the triad Isis + patera, Harpocrates and Anubis, e.g. 
Mus. Naz. Naples inv. no. 19.301 = Tran tarn Tinh 1964, 171 no. 133b pi. 22.2; cf. 
the statue of Isis in the temple at Luxor, which probably once held a patera: Walters 
1988, 112 pi. 52d (c. AD 126); Sarapis + patera: e.g. Nile-fresco from the Casa delle 
Amazoni, Pompeii, showing Isis beside Serapis holding a corpucopiae and a patera: Tran 
tarn Tinh 1964, 127f. no. 13 = Merkelbach 1995, 547 Abb. 66; bronze plaque from a 
villa outside Pompeii, now in the Antiquario, inv. 1090/4, showing Serapis and Isis on 
either side of an altar/base with (a statue of?) Harpocrates standing on it: Tran tarn 
Tinh 1964, 201 no. 105bis; marble relief from Delos, now in Athens, with Isis + situla 
and Sarapis + patera on either side of the Agathodaemon: Dunand 1981 no. 3. The 
well-known addition to the Rabirii funerary relief from the Via Appia (CIL VI 2246 = 
ILS 4404 = RICIS 501/0160), showing Usia Prima, represents a sistrum on one side 
of her head (spectator's left), and a patera on the other: Collezioni egizie, 118 no. 87 = 
Malaise 1972b, 124: Rome no. 35. There are many other examples. 

444 Poo 1986, 1187f; Torok 1995, 62f. 

445 As the 'reed cista', cf. Griffiths 1975, 222-26. For Osiris, see Chap. 2. La above. 
4 "' M. Heerma van Vos, The cista mystica in the Cult and Mysteries of Isis, in 

Vermaseren 1979, 23-26. 

447 Tibullus 1.7.47f; Apuleius, Met. 11.11 with Griffiths 1975, 224; Plutarch, de hide 
39, 366f2. 

448 Altars: e.g. CIL VI 344 = 30744 = Helbig 4 2 no. 1189 = Malaise 1972b, 113 
Rome no. 5 = Lembke 1994, 245, Kat. E49, pi. 46.1 - RICIS 501/0121; CIL VI 13454 
= Malaise 1972b, 127 Rome no. 45 - RICIS 501/0194, a cista on both lateral faces 
of the funerary relief of Balbullia Varilla; AE 1964: 111= RICIS 502/0601 (Cereatae 
Maritimae), cista on front, etc.; Pompeii, north-west wall of sacrarium, between two 
large agathodaemon-snakes: Tran tam Tinh 1964, 143f. no. 47 = Dunand 1981 no. 7 = 



318 CHAPTER FOUR 

very well-equipped Iseum (or Isea) at Beneventum, where the majority 
of the surviving furnishings are in diorite, boasted a cista in porphyry 
a notably hard and expensive ornamental stone from the quarries of 
the Eastern Desert. 449 In all these cases the cista is encircled by a snake, 
and often decorated with a crescent moon. 

Lastly it is worth emphasising the extent to which offerings and sacri- 
fice constructed a coherent world within the Egyptian cults. Such actions 
contributed to a sense of tranquillity on the part of the worshipper, the 
product of a confidence that cultic acts contributed to the maintenance 
of the order of things. The fulfilment of cultic obligations of course had 
such an effect in all contexts, most obviously in the implicit guarantee 
for the general well-being of the community provided by the regular 
offering of civic sacrifice. But if we are to believe Apuleius, this sense 
was particularly marked in the case of the Egyptian cults. There is, for 
example, a clear contrast between the roles of animals, and sacrifice, 
in Books 1—10 of Apuleius' Metamorphoses and that of Book 11: in the 
latter, both animals and humans achieve their proper place in a world 
ordered by divinity 450 With regard to sacrifice, the book conveys an 
impression of sacramental calm, stability and order, by contrast with 
the disorientation, alienation and misunderstandings that pepper the 
preceding books. 451 

ii. Prayer, Healing and Incubation 

Good Isis, hear me this prayer, though thou deny me 

a matter of more weight; good Isis, I beseech thee! 

Charmian in Antony and Cleopatra, Act 1 , scene 2 

Prayer is the established means whereby individuals can approach divin- 
ity and submit requests for help. Apuleius suggests that, in the case of 
Isis, this was not the usual votive transaction, whereby the petitioner 
undertakes to repay the deity if the prayer or request be answered. 



Merkelbach 1995, 507 Abb. 26 (false interpretation); other examples: Tran tam Tinh 
1964 nos. 18 and 67. 

449 Miiller 1969a, 106f. no. 289 = Malaise 1972b, 304 no. 52 = Merkelbach 1995, 
674 Abb. 214. 

450 C.C. Schlam, Man and Animal in the Metamorphoses of Apuleius, in Hijmans 
and Schmidt 1981, 115—142; idem, The Metamorphoses of Apuleius. On Making an Ass of 
Oneself '(London 1992) 100. 

451 T.D. McCreight, Sacrifical Ritual in Apuleius' Metamorphoses, in H. Hofmann 
(ed.), Groningen Colloquia on the Novel 5 (Groningen 1993) 31—61 at 56f. 



THE RITUAL SYSTEMS 319 

Rather, the deity is approached by stressing the asymmetry between 
her omnipotence and the utter humility of the petitioner, expressed in 
the metaphor of servitude. Laid low by destiny, the petitioner throws 
himself on the goddess' mercy {infideni). At least in the idealised version 
of the Metamorphoses, this gesture of abjectness prompts the outpour- 
ing of her benevolence and the appropriate transformation of nefaria 
Fortuna into a Fortuna videns, a Fortune who can see clearly what she is 
about (Met. 11.15). If the Isiast disregards the goddess' commands or 
refuses her offers, he is punished; the punishment ceases only when he 
repents and openly admits his fault, sometimes, as we saw earlier, in 
public. 452 

This pattern of insistence on the asymmetry between god and wor- 
shipper has been familiar since the remarks of A.D. Nock and F. Bomer 
on the phrase kux' znuayr\v or Kara npoaxay\xa, and the notion of 
So'BA.oc; Geov, observations that were very largely confirmed by H.W. 
Pleket. 453 Whereas Nock, and less explicitly Bomer, argued that it was 
implicitly based upon the non-Classical model of the absolute Helle- 
nistic ruler, Pleket urged that we should see the usage as an expression 
of affectivity as well as a mere insistence upon distance. However, he 
agreed that the pattern becomes far more marked in the Hellenistic and 
Roman periods, particularly in the case of Isis, certain Lydian cults, the 
Jewish god, and early Christianity: "The incidental designation of the 
worshipper as 'servant' of the helping, friendly deity in the context of a 
profound personal religiosity and in emergency situations develops into 
a structural phenomenon in the cult of the almighty Oriental deities in a 
markedly hierarchical period" (1981, 165, cf 171). Social dependence 
is represented as pure religious feeling, thus veiling the legitimation of 
the social order effected by the naturalisation of hierarchy in this type 
of religious discourse. Of course such language was, or came to be, 
widely current in the Roman empire, indeed catachrestic: one of the 
commanders at Vindolanda, ] Karus, who addresses one of his cor- 
respondents asfrater — evidently therefore a status-equal — also addresses 
him in the same letter as dominus. 45 * We should therefore at the same time 



152 See Dunand 1973, 3: 216f. and Chap. 3.3 above. 

453 Nock 1925, 95-7 (= 1972, 1: 45-8); Bomer 1957-63, 3: 207f.; Henrichs 1976; 
Pleket 1981. Although Pleket (1981, 155) claims that there are traces of these ideas 
in Classical evidence, so far as I can see he provides none except for the problematic 
case of Ion in Euripides' play (164E). 

151 A.K. Bowman andJ.D. Thomas, Vindolanda: The Latin Writing-Tablets. Britannia 
Monographs Series 4 (London 1983) 105—111 no. 22 1. 5f. rogo ergo domine (the last 



320 CHAPTER FOUR 

recognise the possible limits of such legitimation. 455 "Toute domination 
symbolique suppose de la part de ceux qui la subissent une forme de 
complicite qui n'est ni soumission passive a une contrainte superieure, 
ni adhesion libre a ses valeurs" (Bourdieu 2001, 78). Such language 
conveyed a degree of commitment to the deity that was deemed proper, 
and perhaps little more. 

However that may be, the appropriate place for uttering a prayer was 
the temple, which was deemed the privileged spot for communicating 
with the deity, transmitting requests and obtaining a response by means 
of a vision, a dream or some significant sign of the kind discussed earlier. 
As often as not, it was there that the goddess revealed herself to the 
supplicant. 456 The form, usual in other cults too, was for the prayer to 
be uttered over a burning altar. 45 ' Women loosened their hair to pray 458 
One might also kneel, as documented by some of the naophoros statues 
from Beneventum, Ostia and the Iseum Campense, and commonly in 
Egypt; but worshippers are often described as seated, no doubt with 
their eyes fixed on the statue. 459 Apuleius described how Lucius prays 
to the goddess' statue early in the morning, when the doors have been 
opened and the white curtains drawn back to reveal the image. 460 
People prayed for the ordinary favours, health and recovery of health, 



four letters are dotted); cf. in 1. 4: the reading domijne is reasonably secure; 126—32 
no. 37 11. 4-6 with p. 92 (on no. 4 1. 36. The editors cite P.Hib. 276.3H (= CPL 260): 
et praese(n)s te domine Jrater . . . ; and P.Oxy 32.4H (— CPL 249): peto domine ut eum...\ by 
the third century, dominus evidently might just mean 'my dear chap': PDm: 63 = Fink 
1981, 355 no. 88 recto 1. 5: peto domine frater .. .; PDur. 64 = Fink 1981, 365 no. 89 12 
col. i 1. 7: opto te djomine. . . 

455 Note the implicit reserves of Nock, 1928, 83f. (= 1972, 1: 74f.). 

456 E.g. Apuleius, Met. 11.5: En admm tuis commota, Luci precibus...; cf. Malaise 1980, 
113. 

457 E.g. Vergil, Aen. 4.219; other forms of respect and admiration, not necessar- 
ily linked to prayer: Cicero, 2Verr. 4.94 (embracing the statue too); Ovid, Met. 9.772 
(embracing altar); Apuleius, Met. 4.28 (putting thumb and forefinger to the mouth). In 
general, see Appel 1909, 190f; 199f; C. Guittard, Ritualisme et sentiment religieux 
dans la priere a Rome et en Ombrie, in A. Caquot and P. Canivet (eds.), Ritualisme et 
vie interieure. Religion et culture (Paris 1989) 19. 

458 Tibullus 1.3.31: resoluta comas, and 90: longos turbata capillos, with Smith ad loc; 
Ovid, Met. 9.772: passis. . . capillis, with Bomer ad loc. 

459 Kneeling: Miiller 1963a, 104 no. 287; 108 no. 287; 290; Lembke 1995, 231 no. 
E25, pis. 38.1-2 (XXVIth Dynasty— Egyptian import); sitting: Tibullus 1.3.30; Ovid, 
Amores 2.13.17: saepe sedit. 

460 Met. 11.20: velis candentibus reductis in diversum, deae venerabilem conspectum appre- 



THE RITUAL SYSTEMS 321 

protection in child-birth, 461 for the family, particularly children, 462 safe 
return from a journey, escape from poverty 463 success in business or 
profession. 464 As Nock emphasised: "Isis and Sarapis were commonly 
thought of as deities effective in an ordinary way". 465 

Prayer of this type is occasionally linked to miracle. A famous fictional 
case is Ovid's narrative of the transformation of Iphis into a man. 466 
Lygdus, the husband of Telethusa of Phaestus, had instructed her to 
expose a female child if one should be born to her, but Isis appeared to 
the expectant mother in a dream and bade her rear the baby, quidquid 
erit (1. 699). With the connivance of the wet-nurse, the girl was brought 
up as a boy named Iphis after the grandfather. In due course, Lygdus 
arranged a marriage with Ianthe; in her desperation for the decep- 
tion not to become public, with all the concomitant scandal for the 



11)1 There are a number of votives from northern Greece to Isis Ao^ta, e.g. SEG 34 
(1984) 627 - RICIS 1 13/0218 (Dion); IG X.2, 97 = 1 13/0523 (Thessalonike). Despite 
the minimal votive evidence (none of the dedications to Bubastis mention the grounds 
for the dedication), there is good reason to think that Isis-Bubastis had special care for 
confinements: Ovid, Met. 9.691; Nicarchus ap. AnthPal 11.18 (mid-P), with F. Dunand, 
Une interpretatio romana d'Isis: Isis, deesse des naissances, REL 40 (1962) 83-6; 1973, 
3: 263-5; eadem, s.v. Boubastis, LIMC 3 (1986) 144f; Malaise 1972a, 189f; Heyob 
1975, 70-3; G. Wagner, Une nouvelle dedicace a Boubastis, ASAE 69 (1983) 247-52; 
J. Quaegebauer, Le culte de Boubastis-Bastet en Egypte greco-romaine, in L. Delvaux 
(ed.), Les divins chats d'Egypte: un air subtil, un parj'um dangereux (Leuven/Louvain 1991) 
117-27. Isis enabling conception: Isidorus, Hymn 2 1. 15f. (p. 36 Vanderlip) - Totti 
no. 22. J. Bergman, Isis-Seele und Osiris-Ei: Zwei dgyptologische Studien zu Diodorus Siculus 
1.27.4-5 (Uppsala 1970) argues that the connection between Isis, Bastet and procreation 
was established through the Egyptian New Year festival. 

11)2 Cf. V von Gonzenbach, Untersuchungen zu den Knabenweihen im Isiskult der romischen 
Kaiserzeit (Bonn 1957); Malaise 1972a, 150f; 1986, 110; Dunand 1973, 3: 170; Tran 
tarn Tinh 1973, 22f; Heyob 1975, 77f; J. Alvar et al., La religiosidad misterica en el 
espacio familiar, ARTS 1 (1998) 218—29. For two further images of children devoted 
to Isis, see F. Baratte, Deux portraits d'enfants isiaques, RA 1993, 101—10 (Mauretania 
Caesariensis); also the well-known early IV P funerary stele from Ostia, found in the 
Fascist excavations: Malaise 1972b, 86 no. 117 with pi. 9. 

*63 xhis is the purported occasion for the sacrifice described in AnthPal 6.231 — 
Garland, Philip no. xxn. 

"'' Influence with the King, fame and good health as a result of building a Serapeum: 
P^enon 59034 1. 19—21 = Totti no. 71; success in trade: Isidorus, Hymn 2 1. 5f — Totti no. 
22: oooot Got £'i5x ov ' cal eit' Euitopuyv xe TtapeTvcu, jiXowoCo' etxjefSeei; si? xov cbiavxa 
Xpovov; Osiris promises Lucius success as an advocate: Apuleius, Met. 1 1.27: nam et illi 
(i.e. Lucius) studiorum gloriam et ipsi (i.e. the pastophorus) grande compendium sua (i.e. Osiris') 
comparari providentia; 30: liberali deum procidentia iam stipendiis forensibus bellule fotum. 

""" He continues: "...The rhetorician Aristides was deeply devoted to Sarapis as a 
god able to save from sickness and shipwreck and to work other miracles...": Nock 
1932, 355 (= 1972, 1: 304). 

166 Ovid, Met. 9.666-797, cf. Dunand 1973, 2: 83: "II n'y apparait rien qui soit 
specifiquement cretois". 



322 CHAPTER FOUR 

family, Telethusa goes to the temple to pray to Isis for help. The goddess 
replies by shaking her altars and causing the temple-doors to rumble; 
then the miracle occurs: nam quaefemina nuper eras, puer es (1. 790f), and 
is — rather oddly given the situation — properly documented by a votive 
tablet: Dona puer solvit quaefemina voverat Iphis (1. 794). 467 The truth of the 
narrative is underwritten by Ovid's implied theoria. v& 

Perhaps the most characteristic mode of intervention by Isis and 
Serapis was their protection of sailors, an area of activity that was both 
grounded in and ratified by the main spring festival of the Isidis navigium. 
One of the Oxyrhynchus papyri contains a fragmentary account of a 
miracle narrated by the helmsman Syrios. 469 As reconstructed by Otto 
Weinreich, the story runs as follows: the narrator's ship, having been 
saved from wreck by Serapis, ran out of fresh water; but Syrios let a 
bucket down into the sea, drew it up, found it was sweet, and attributed 
the miracle to Serapis. 470 After an intervention by an unknown party 
(possibly a manifestation of the god) the miraculous water was sold to 
the people of Pharos, and an account of the miracle (6cpexr|) depos- 
ited in the temple of Thoth-Mercury The by-standers gave thanks: oi 
7iap6vx£c; zinaxz "etc; Zevc; Zapamq". 

The hymn to Isis from Andros recounts how Isis 

7tpdra 8' otI aeX|iati 8o{)p(ov 
KoA,7icoxav oGovatat Ooav Tp07uv iS-uvecKov 
oi8|ia KaQinnevoiaa, 8a|ia^o|ieva<; 8e Qakaaaac, 
coKi)7t6poi<; eXaxaiq eXtKocv eaxaae xopeiav 
A(opi8o<; ei)Xoxia. 471 



467 This is the sole version of the Telethusa story to survive, but Nicander knew an 
identical miracle by Leto (Phytia), in favour of Galatea, wife of Lampros of Phaes- 
tus, and her daughter/son named Leucippus: Antoninus Liberalis no. 17. Nicander 
explained the name of the festival 'EKoucua by this means: tov niitXov f| jtccu; e^e8t); 
cf. G. Weicker, s.v. Galateia 2, RE 7 (1910) 518f; Nilsson 1906, 370f. 

468 I. Rutherford, Tourism and the Sacred: Pausanias and the Traditions of Greek 
Pilgrimage, in S.E. Alcock et al. (eds.), Pausanias: Travel and Memory in Roman Greece 
(Oxford 2001), 40-52 at 42f. 

469 POxy 1382 = Totti no. 13; the papyrus is dated IP. 

4/0 See Weinreich 1919, 13-18; Aelius Aristides, Or. 45.29: oxixoq, t)8(flp dvfJKE Jt6ti)j,ov 
ev |ie<yp %aX6.i%\\, alludes to the same miracle. The waters of the Nile, especially at the 
time of the flood, flow a considerable way into the Mediterranean. 

471 "Standing on the deck, I was the first to guide responsive ships, their sails swelling, 
over the sea; the glorious progeny of Doris has staged a plunging dance on the sea 
tamed by swiftly-travelling vessels . . . : IG XII.5, 739 - Totti no. 2 = RICIS 202/1801, 
11. 152—6, §50). The date is uncertain, but probably late P. 



THE RITUAL SYSTEMS 323 

This myth-history asserts a norm, that the waves of the sea shall be 
mastered by ships' hulls, and implies that Isis will ever be at hand to 
ensure that it is so; at the same time, the subtext implied by double 
reference to danger from wind and waves reminds the reader of the 
constant uncertainty of journeying by sea (Dunand 1973, 2: 117f.). 
Rather more than a century later, in his proempticon for M. Maecius 
Celer (cos. suff AD 101), Statius composed a literary prayer for the 
officer's voyage from Puteoli to the East: 

Isi, Phoroneis olim stabulata sub antris, 
nunc regina Phari numenque Orientis anheli, 
excipe multisono puppim Mareotida sistro; 
ac iuvenem egregium . . . 
ipsa manu placida per liminafesta sacrosque 
due portus urbesque tuas. 472 

Here the two injunctions, excipe and due, express the wish that the deity 
pay personal attention to the ship's fate: the rattle of the sistrum is to 
scare away misfortune; the accumulation of positive adjectives, manu 
placida, liminafesta, sacros portus, alludes to the common phraseology of 
prayers for successful outcomes, while artfully evoking the milling crowds 
in the harbour of Alexandria (or Caesarea Maritima). 473 The goddess 
did not dispense acQTiptcx indiscriminately: her aid was focused upon 
particular situations and emergencies, and of course the uncertainties 
of the deep sea meant that there were many occasions when her help, 
like that of the Great Gods of Samothrace, could be invoked. 474 In 
return, grateful sailors offered votive-plaques illustrating the storm they 
had survived. 475 Who wanted to be reminded that her failures could 
hardly reproach her from the depths? 

Another type of danger was sexual violence. As I pointed out in the 
previous chapter, female chastity was a central value in the Egyptian 



472 "Isis, once stabled in Inachus' caverns, now Queen of Pharos and deity of the 
sultry Orient, take (this) Egyptian vessel under the protection of your rattling sistrum 
and with a peaceful hand guide this eminent young man yourself to harbours decked 
for rejoicing and to port-cities under your special protection" (Silvae 3.2.101-7). Celer 
(PIR 2 M 51) had been appointed legionary legate in Syria by Domitian, probably in 
AD 94. On the cult of Isis in Statius, cf. S. Montero, Los dioses egipcios en Estacio, 
Habis 10-11 (1979-80) 241-53. 

4/3 Placidus first occurs as an epithet of gods in the Augustan period; the adjectives 
evoke the common wish quod bonum faustum felixque sit, cf. Hickson 1993, 58f; 62-83. 

4/4 E.g. the casual mention in AnthPal 6.231 1. 7 — Garland, Philip no. xxn of being 
saved from the sea: &<; 8K TieXdyoix; eppuaao (i.e. Isis) A&ruv. 

1/3 Juvenal, Sat. 12.27f: pictores quis nescit ab Iside pasci? (cf. n. 412 above). 



324 CHAPTER FOUR 

cults, and Isis' role in inventing and maintaining marriage is an impor- 
tant theme in the aretalogical literature. 476 It is therefore not surprising 
to find her being appealed to in Xenophon's Ephesiaka}' 1 When, after 
numerous adventures, Anthia is taken by the robbers to Memphis, and 
Polyidos, after defeating them, besets her with importunities, she eludes 
him and runs to the temple of Isis: av |ie, eiTtev, cb 8eo7ioiva Aijvnxov, 
7tdA.iv ocooov, fi e(3or|9r|oa(; 7toA,A,dia<;- cpetadaBco iiov> Kai noA/uiSoq xr\q 
Sid oe oaxppovo!; 'AppoKouri xr|poi)H£vr|<;. 478 Polyidos then swears to 
respect her chastity, but takes her to Alexandria, where his wife abuses 
her and sells her a slave to a brothel-keeper in Tarentum. 479 Later in 
the same book, the two lovers, now united, go to the temple of Isis in 
Rhodes and thank the goddess for her aid in bringing them safely, i.e. 
without the loss of their sexual purity, together once again: Sid oe, cb 
Ttdvxcov fiLuv xiLiicoxdxri, kavxovc, d7t£iA,r|(pocLi£V (5.13.4). The central 
place of marital fidelity in the imaginary civic world of the novel is 
summarised by the narrator's claim that the search for Anthia was onjxcp 
xov pico Ttcxvxoq Kai xfjq 7tA,dvr|q f) vnoBzaic,, the basis of Habrocomes' 
entire life and all his journeying 480 Although this is obviously a fictional 
exaggeration, Swain is surely right to see conjugality as one important 
aspect of the values of the civic elites of the eastern Mediterranean, 
which helped to guarantee the viability of the oik6<; both as a private 
institution and as the basis of the public life of the city (1996, 127f) 
But sometimes prayer was offered without immediate gain in mind, 
as a mark of repentance or simply as disinterested praise. In such cases, 
prayer approaches its pure religious form, as total surrender to the deity, 
a true deditio infidem, without any ulterior motive. The early encomium 



476 See Chap. 3 n. 133 above; also Malaise 1972a, 169; Dunand 1973, 3: 261-5; 
Heyob 1975, 45-52; 66-80. 

477 The lucubrations of Merkelbach 1962, 72-90, extrapolated from Kerenyi 1927, 
are entirely fanciful; cf. Swain 1996: 108 n. 21: "The novels' interest in religious mat- 
ters simply reflects the pervasive nature of religion in ancient society at all periods". 
Isis appears as the goddess of the marriage -bond in Achilles Tatius, Kleitophon and 
Leucippe 5.26.4. 

478 Ephesiaka 5.4.6; cf. Merkelbach 1995, 361-3. 

479 In an analogous earlier incident, Anthia claims to Psammis that at her birth her 
father had dedicated her as a iepa ttj<; "IaiSoi;, which seems to mean that she was 
vowed to chastity until she was married (3.11.4), cf. Merkelbach 1962, 104; Dunand 
1973, 3: 170. 

480 5.8.2. The fictional date of the story is vaguely 'Hellenistic'; the novel may have 
been written between AD 100—150, but I rather incline to Dalmeyda's scepticism that 
it is datable at all. 



THE RITUAL SYSTEMS 325 

of Isis from Maronea in Thrace stands out in this connection. 481 The 
author had evidently had his eye-sight restored by the goddess (cocmep 
ovv etc! xcov ouudxcov, 'lot, xcfic, evxaic, [e.nr\K]ovaac,...\. 6f). Since 
the goddess attended the speaker when he implored help from her, a 
fortiori she will attend now that his only intention is to praise her (n5>q 
vnzp xr\q iSicxc; xtufjc, ovk otv eA,9otc;; 1. 1 1). He has every confidence 
that he will be heard, since an encomium may be penned by human 
hands but is in fact inspired by god (yiyvcooKcov oxi x6 eyiccoruov, vov>c, 
|o,ev Geco, xeTpec; 8e ypcxtpcuaiv avBpomoi) . . . 1. 12f.). 482 

The phraseology of Lucius' prayer while still an ass at the beginning 
of Book 1 1 , and that of his thanksgiving some chapters later, after he has 
regained human form, are likewise clearly modelled upon the pattern 
of the encomium. 483 The second passage concludes with the assurance 
to the goddess of constantly-renewed personal devotion: 

quod solum potest religiosus quidem, sed pauper alioquin, efficere curabo: divinos tuos 
vultus numenque sanctissimum intra pectoris mei secreta conditum perpetuo custodiens 
imaginabor. m 

Met. 11.25 

A similar idea is to be found in the poem of Damaios at Thessa- 
lonike, already mentioned, where the speaker asserts that if Osiris, as 
requested, sends good reputation and good health to Phylakides and 
his two sons, he will have taught mankind the lesson of piety: otppoc 
xiq &u£p{cov Xevoocov xdSe 9i)ii6v 6xpwr|i atpcoixepoiJ uaicdpcDLi nr|7ioxe 
Afjaxw exevv. 483 



481 SEG 26 (1976) 821 - Grandjean 1975, 17-21 = Totti no. 19 = RICIS 1 14/0202 
(F). The encomium of Serapis by Aelius Aristides, Or. 45.16-32, is calqued upon such 
compositions. 

182 Cf. Grandjean 1975, 38—44 on the motif of divine inspiration in the classical 
tradition, e.g. Pindar, Pyth. 79-82. O. Weinreich, Hymnologica, ARW 1 7 (1914) 527-30 
collects some apt passages from Aelius Aristides, e.g. Or. sacr. 4.31 (vol. 2, p. 433 Keil; 
tr. Behr, p. 324) f|X9e ydp not evvraviov ippd^ov tov xe naiava wq §eov jtoifjocu tib 9eS 
Kal 6i|Ja xriv dpxijv otuxoi) . . .; 39 (p. 435 K., tr. Behr, p. 326): t\kev Se Kai nap' 'Afrnvat; 
ovap iVvov e'xov zr\q Qtoxi Kal dpxriv -coidvSe 

483 Met. 11.2; 25. 

484 "I shall therefore take care to do the only thing that a devout but poor man can: 
I shall store your divine countenance and sacred godhead in the secret places of my 
heart, forever guarding it and picturing it to myself" (tr. Hanson). 

485 "So that a man may behold this and urge himself never to forget the gods, 
who have no cares": IG X.2, 108 11. 9f. = Totti no. 72 = RICIS 113/0506; cf. n. 379 
above. 



326 CHAPTER FOUR 

Here the emphasis is clearly upon the |xr|7i:oxe, on the continued liv- 
ing presence of the god in the heart. 

The hymns of Isidorus from Medinet Madi in the Fayyum provide 
another good example of the type of encomium offered Isis. In Hymn 
1 , after listing some of her inventions and gifts of nature (light, winds, 
the Nile flood), and surveying the foreign peoples who glorify her name, 
he concludes: 

Mighty one, I shall not cease to sing of Your great Power, 

Deathless saviour, many-named, mightiest Isis, 

Saving from war, cities and all their citizens: 

Men, their wives, possessions, and children. 

As many as are bond fast in prison, in the power of death, 

As many as are in pain through long, anguished, sleepless nights, 

All who are wanderers in a foreign land, 

And as many as sail on the Great Sea in winter 

When men may be destroyed and their ships wrecked and sunk . . . 

All (these) are saved if they pray that You be present to help. 

Hear my prayers, O One whose Name has great Power; 

Prove yourself merciful to me and free me from all distress. 486 

Hymn I, 11. 25-36 (Bernand 1969, 633) = Vanderlip 1972, 

17f. = Tottino. 21 

In view of my earlier discussion (see chap. 3.3) of the aretalogy as a 
genre, it is unnecessary to pursue this issue further. It is more worthwhile 
to look at some individual practices. One custom known from Egypt 
is the proskynema, adoration of the deity, which was a term denoting a 
pilgrimage to the temple in order to adore the deity 487 An early example 
from the temple of Isis at Philae is a pilgrimage performed by an 
extremely important official of the Ptolemaic court, the e7iioxpaxr|y6<; 
Callimachus II, governor of Upper Egypt, the Red Sea and the Indian 
Ocean, in the mid-first century BC, on behalf of the King, Ptolemy XII 
Auletes: tJkco rcpoq xf|V lcopiav 'low koc! 7tejx6r|Ka xo rcpoaicuvrpa xov 
Krjpiorj pocaiAicoq Beot} varo Aiov6<o>otj <lHA,o7xdxopoc; OiAcxSeAxpoi). 488 



486 -p r Vanderlip; part of this section was cited earlier in relation to miracle, n. 100 
above. Cf. also Hymn 2, 5-8: "All who pray to you. . .if they (but) pray to you, (they) 
quickly obtain renewal (of life) from you"; 1. 29f: "Grant a share of your gifts also to 
me, Lady Hermouthis, your suppliant ". 

187 The evidence for one aspect of such pilgrimage, Isiac footprint-votives, is col- 
lected by Takacs 2005, 361-6. 

488 OGIS 186 = SB 4084 = Bernand and Bernand 1969, 1: 306-11 no. 52 = Totti 
no. 27. On the man, and his father, see briefly Fraser 1972, 1: 182; 2: 189f n. 82; on 
this type of proskynema, ibid. 2: 315 n. 401. 



THE RITUAL SYSTEMS 327 

This approach was presumably on behalf of the entire kingdom. At a 
lower social level, a common intention was to obtain a blessing for the 
whole family. A simple, and fairly ungrammatical, example, written on 
another of the pylons of the temple of Isis at Philae (where dozens of 
such messages have been found), reads: to 7tpoaic6vr||icx 'Apidviv xov 
Kai 'AtcoAAcoviov kou xr\q awfKoi) a\)xo\) Kcxi xcov xekvcdv icai xov oikoi 
b"kov napa xr\q rropiovoVoi) "IaiSoq, or||j.epov, in' ayaGcp, (exovq) A-oc, 
OanevroG k0. 489 

Such pilgrimages might indeed be repeated over years: a series of 
late texts from the memorial temple of the Female King Hatshepsut 
at Deir el-Bahari (Western Thebes) records irregular visits made over a 
thirty-year period by a sodality of iron-workers (7tA,fj9oc; oiSriponpycov) 
from Hermonthis to offer sacrifice and adore the god. The organiser 
seems to have been a professional scribe-cum-priest named Hatres son 
of Horion; the events took place over two days, including an overnight 
stay in the temple precinct and no doubt a ritual banquet in the late 
afternoon. The best-preserved dipinto specifies that the animal offered to 
the god was a donkey, a typical victim in Egypt in honour of the great 
gods, though virtually unknown in Greece and Italy, where the animal 
was considered unsuitable for sacrifice: avxbq (the donkey-keeper Plenis) 

eo<pa^ev xov ovov euTtpoaBev xov Gecu Kai navxzq xb Jtpoaic6vr|Utt xov 

ueydAxn) Geov znoi[r\aav . i90 One of the characteristic features of these 
texts is the wish in' dya0cp; another is the participle |ivr|o9e{(;. 491 Both 
also occur in the context of the Egyptians cults in the Mediterranean, 
for example at Delos, Pompeii, Ostia, Rome, and in North Africa, 
indicating that the practice of adoration-pilgrimage was known there 
too. The graffito at Pompeii, for example, records the proskynema of a 
man from Beroia, no doubt in gratitude for a safe voyage: e|j.vr|o9r| 
QeotpiAoc; Beporiq in' dyocBcp napa xfj Kvpia. m 



489 '"phe proskynema of Harkinis, also known as Apollonios, and his wife and his 
children, and their entire household, before Isis of the many names, today, 31st year 
(of Commodus), 29 Phamenoth (i.e. late March, AD 191)": SB 8681 = Bernand and 
Bernand 1969, 2: 166-74 no. 168 = Totti no. 35 11. 14-19; on the finds at Philae, cf. 
Dunand 1973, 1: 150-9. 

490 SEG 41 (1991) 1612 11. 8f., dated Dec. 27/8th AD 324. 

491 E.g. the proskynema of Sarapias at Philae, AD 4: tJko){i} npbq |reydXr|v Eiatv Tf|V 
ev <I>i>Vai<; nvelav in' ayaQSn twv yovecov rcoioiVevoi; Kai iwv aSeXipmv Kai (piXcov not) 
Kax' ovoua ... Bernand and Bernand 1969, 2: lllf. no. 151 11. 2-4 — Totti no. 30. 

492 CIL IV 4289 - Tran tam Tinh 1964, 178 no. 156 = Malaise 1972b, 266 Pompei 
no. 26 = RICIS 504/0215. There must have been a civic sanctuary of Isis Lochia at 



328 CHAPTER FOUR 

The second practice to be noted here is healing. 493 Since the main vec- 
tor of disease in Egypt was considered to be a pathogenic or miasmatic 
movement of the air termed dhrt {deherei), 'bitterness', brought about by 
the messengers of Sekhmet, healing was a very important aspect of 
the services provided by Egyptian temples to their community 494 The 
specifically religious methods used hardly differed from those employed 
by 'secular' and the many types of free-lance practitioners: in the Late 
Period they mainly focused upon drinking water that had been poured 
over certain statues or inscribed stelai of the gods, particularly Horus 
as shed ('saviour', but also 'reciter of texts') — a famous example is the 
Metternich stele now in the Metropolitan Museum, whose obverse 
shows Horus with Isis and other gods. 493 The water was supposed 
to pick up the magical force of the texts it washed over. Many such 
stelai and statues are equipped with a shallow basin in front for col- 
lecting the water thus fortified. At Dendera, and probably in many 
other late temples, patients were expected not merely to drink but to 
immerse themselves entirely. Cures, and protection from dangers such 
as scorpion-stings and snake-bites, could also be obtained by reading 
sacred texts, or touching the papyrus on which they were written. 496 
Complementary to this 'magical' healing, a variety of surgical inter- 
ventions was undertaken, and a complex pharmacopoeia applied to a 
great variety of lesions and ailments. 49 ' In the Late Period, there were 
also at least two specialist healing-gods, known to have been illustrious 



Beroea in Macedon, cf. RICIS 113/0301; Bricault 2001, 23 Map VII with comm. on 
p. 24; Dunand 1973, 2: 190. 

493 Griffiths 1975, 166; Malaise 1980, 108£; Frankfurter 1998, 46-52; Dunand 
2006. 

494 Cf. Etienne 2000, 56—9. On the overlap of the roles of priest, magician and healer 
in Egypt, for example the kherep Serket in the Late period, see Koenig 1994, 30—2. 

493 See e.g. L. Kakosy La magia nell'antico Egitto, in AA.VV, La magia in Egitto 
ai tempi dei Faraone. Mostra Milano 1985 (Modena 1985) 7-101 at 59-66; Konig 1994, 
100-12. Horus as shed: e.g. EHarris VII.12-VIII.1 (p. 64f. no. Q,ed. H.O. Lange). 

496 Statues representing illustrious persons might show them holding such stelai or 
texts (e.g. the 'statue Tyszkiewicz' in the Louvre: Etienne 2000, 111 no. 207, fig. on 
p. 66); placed in a temple, these images would in turn be effective for protection and 
healing. 

497 The standard work is W. Westendorf, Handbuch der altdgyptischen Medizin (Leyden 
1999); cf. also B. Ebbell (ed.), Die alt-dgyptische Chirurgie (E Smith and P Ehers) (Oslo 1939); 
M. Helbling, Der altdgyptische Augenkranke, sein Artzt und seine Gotter (Zurich 1980); W.M. 
Pahl, Altdgyptische Schddelchirurgie (Stuttgart 1993). On the pharmacopoeia, apart from 
Westendorf: R. Germer, Katalog der altdgyptischen Pftanzenreste der Berliner Museen (Wiesbaden 
1988); eadem, Die Heilpflanzen der Agypter (Zurich 2002); T. Pommerening, Altdgyptische 
Heilpflanzen (Stuttgart 2005) (popular). 



THE RITUAL SYSTEMS 329 

doctors in former times: Imhotep-Imouthes, 'son of Ptah', and Amen- 
hotep-Amenotes, who healed the sick at Deir el-Bahari in Western 
Thebes (on the upper terrace of the Hatshepsut-temple) and at Philae. 
Both were identified with Asclepius in the Graeco-Roman period. 498 
Imhotep-Imouthes also possessed a temple close to the Serapeum in 
Memphis/ Sakkara. Demetrius of Phaleron, the pupil of Theophrastus, 
vouched that he had been cured of blindness by Serapis at his shrine 
at Canopus in the Delta. 499 

Although the aretalogies are conspicuously silent about her powers 
in this area, there are many Dynastic-Egyptian Isis-texts for healing; 500 
at her shrine at Dendera there was a sort of 'hospital' where immer- 
sions took place; and, at any rate in the popular religion of Alexandria 
in the Ptolemaic period, she was closely associated with the protection 
of family, childbirth and children. 501 As we have seen, Isidorus' Hymn 
II expressly comments on her ability to heal the seriously ill, 302 and in 
the late first century BC her role as healer is attested by a well-known 
passage of Diodorus: 

Standing above the sick in their sleep she gives them aid for their diseases 
(koto xovq vnvovq etpiaTOiaivnv 8i86vat xdiq Kaiivoixn PonSruiara 7tpo<; 
xac, voaovq) and works remarkable cures upon those who submit them- 
selves to her (xovq vnaKoiaavxaq ai)xf\ napab'oc^aq -uyid^eoSai); and many 
who have been despaired of by their physicians because of the difficult 
nature of the malady are restored to health by her, while numbers who 
have already lost the use of their eyes or of some other part of their 
body, whenever they turn for help to this goddess (orav 7ipo<; rawnv xr\v 
0e6v Karacpnycocnv), are restored to their previous condition. 503 



498 Numerous dipinti in Greek were found in the shrine, expressing gratitude to 
Amenotes (always named Asclepius), cf. F. Bataille, Les inscriptions grecques du temple de 
Hatshepsout a Deir-al-Bahari (Cairo 1951); the best treatment is now Lajtar 2006. The 
best-known is that of Athenodorus, tesserarius of the "First Vexillation" of auxiliary 
troops, who was able to enter the inner cella and there come face to face with Asclepius' 
and be healed: Bataille, ibid., 85-9 no. 126 = Totti no. 17. 

199 Diogenes Laertius, Vit. phil. 5.92, cf. von Arnim, s.v. Demetrios no. 85, RE 4 
(1901) 2817-41 at 2834f. 

500 E.g. J.F. Borghouts, Ancient Egyptian Magical Texts (Leyden 1978) 48f. no. 80; 59f. 
no. 90; 74f. no. 102 etc.; PHarris VII 8-12 (p. 61f. no. P ed. H.O. Lange). Isis' associa- 
tion with magic is doubtless also relevant, cf. Dunand 1979, 1 13—19. 

501 The association with the family was of course stimulated by the motif of Isis 
suckling Horus the Child, cf. Dunand 1973, 1: 95f. 

502 Isidorus, Hymn II 1. 7f. (p. 35 Vanderlip). 

503 Diodorus Siculus, 1.25.5 (tr. C.H. Oldfather); cf. Aelius Aristides, Or. 26.105; 
27.39; 36.124 (Isis and Serapis together). 



330 CHAPTER FOUR 

The Egyptian gods seem to have been adopted fairly widely as 
healing-gods in the eastern Mediterranean. A number of votive gifts 
of body-parts appear in the Delian treasury-lists, which clearly imply 
that healing was conducted both in the Iseum and in Serapeum C. 504 
I have alluded above to the story of Lenaeus and his blinded charger, 
which was healed by Serapis. 505 The late third-century AD — or even 
early fourth- century — Selbstqffenbarung of Karpokrates from Chalkis on 
Euboea, unlike the earlier analogous texts, explicitly claims that the god 
invented the use of simples: naoav (pcxpu«K£{cxv iaxpoiq eiq ocoxripiav 
[e.g. xcov av9pamcov e7tevor|ooc ] Tevtavioq, 'EjuSoa>pio<;. 306 The best- 
known case is of course that of Aelius Aristides, who devoted the five 
years of his life prior to his 'conversion' to Asclepius to intense worship 
of Serapis, and indeed renewed his commitment when he returned 
to Smyrna: "Of all the gods whom Aristides worshipped, aside from 
Asclepius, none held a more important place in his life and belief than 
Sarapis" (Behr 1968, 149). 

The statement by Diodorus raises the question of whether incuba- 
tion (eyKo(ur|oi(;, incubatio) was known in the Graeco-Roman cult of 
Isis and Serapis, and if so, how widespread it may have been. 50 ' We 
must of course distinguish here between the two types of incubation 
conventionally recognised, therapeutic and divinatory (Renberg 2006, 
105). As my sketch of Egyptian healing-practices suggests, incubation 



504 E.g. two eyes on a plaque, uninscribed; some eyes in gold, given by C. Mes- 
sius; a golden eye given by Epiteugmas; a pair of eyes in relief on a pinax, given by 
Demetria; numerous other eyes; some fingers, by P. Aemilius; some genitals; an ear; 
a silver throat; two small feet; etc.: ID 1417 = RICIS 202/0424; ID 1452 = RICIS 
202/0433. From the useful index in R. Hamilton, Treasure Map: A Guide to the Delian 
Inventories (Ann Arbor 2000) 465—79, it seems that such objects were offered only to the 
Egyptian gods on the island. 

505 Aelian, JVA 11.31, cited p. 180 above. 

506 Harder 1944, 7-18 = Totti no. 6 = RICIS 104/0206 1. 11; cf. Dunand 1973, 2: 
152—4. 'Karpokrates' is a late rendition of Harpokrates; the text was compiled by a 
local rhetor, drawing upon the encomium/aretalogical topoi. The epithet 'E7ri5aijpi0(; 
is likewise unique in the genre, and clearly reinforces the medical claim; Teitavioi;: a 
glancing reference to Titane near Sicyon, famous for its temple of Asclepius (Pausanias 
2.11.5-8; 23.4; 7.23.8, cf. Herondas 4.14-16; Edelstein 1945, T.555 and 536). 

307 L. Deubner, De incubatione capita quattuor (GieBen 1900); N. Lewis, The Interpretation 
of Dreams and Portents (Toronto 1976); A. Martin-Artajo, En torno a la incubatio, in Alvar, 
Blanquez and Wagner 1994, 135—43. In the context of initiation and thence salvation: 
M.J. Hidalgo, Los oraculos y los suenos-visiones como vehiculos de salvation en las 
novelas greco-romanas, in Alvar, Blanquez and Wagner 1992, 175—204. 



THE RITUAL SYSTEMS 331 

in the Greek sense was not traditionally practised there. 508 However, 
the equation of Imhotep-Imouthes with Asclepius in the Ptolemaic 
period seems to have led to a limited introduction of Greek therapeutic 
incubation, at least in the Delta area. Strabo at any rate confirms that 
incubation for cures, both on one's own account and for others, was a 
feature of the 'very holy' temple of Serapis at Canopus in the western 
Delta. 509 There was indeed a regular genre of dreams that resulted in 
genuine cures: a certain Artemon of Miletus filled twenty-two books 
with them, many of them figuring Serapis. 510 Aelius Aristides, who 
repeatedly invokes the healing powers of the Egyptian gods, specifies 
that these were performed, as in the case of Asclepius, through divine 
visions, Geicxq o\|/eiq. 511 

The influence of the model of the Epidaurian m^iaxa on healing 
practice in Roman Egypt — or at any rate on its temple-propaganda — is 
suggested by a first-person miracle-narrative written on the verso of the 
Oxyrhynchus papyrus in the Bodleian that contains the encomium of 
Isis, dating from the first half of the second century AD (POxy 1380). 
Col. v contains the following narrative: 

It was night, when every living creature was asleep except those in pain, 
but divinity showed itself the more effectively (to 8e 9eiov evepyeatepov 
ecpaivexo, 1. 94f.); a violent fever burned me, and I was convulsed with 
loss of breath and coughing owing to the pain proceeding from my 
side. Heavy in the head with my troubles, I was lapsing half-conscious 
into sleep, and my mother, as a mother would for her child . . . was sit- 
ting without enjoying even a short period of slumber (|ir)8e koc9' oWyov 
i)7tvoi) |ietava^a|ipdvo\)oa, 1 06f), when suddenly she perceived — it was 
no dream or sleep, for her eyes were open immovably, though not seeing 
clearly (6(p0ocA,|ioi ydp r\aav aKeivr|toi 8ir|voiy|xevoi, fiXinovxeq |iev o\)k 
ctKpiPttx;, 1. 109—12) — a divine and terrifying vision, easily preventing 
her from observing the god or his servants, whichever it was. In any 
case there was someone whose height was more than human, clothes in 
shining raiment and carrying in his left hand a book, who, after merely 
regarding me two or three times from head to foot, disappeared. When 



508 Oracular incubation, analogous to katarchic astrology, was however a common 
feature of Egyptian religious practice; cf. J.D. Ray, The Archive of Hor. Texts from 
Excavations 2 (London 1976); Depauw 1997, 150. 

509 Strabo 17.1.17, 801C: kou eyKoi|xaa9ca ocoxotx; •unep eaoTcbv f) e-tepoix;; Artemi- 
dorus, Oneimcr. 4.22 stresses the number of people who had been cured by dreams at 
Alexandria. See also Dunand 1973, 1: 112f.; 3: 141; Dunand 2006, 10. 

510 Demetrios: Diogenes Laertius, Vit. philos. 5.76; dream-books: Artemidorus, 
Oneimcr. 2.44. 

511 Aelius Aristides, Or. 47 (= Sacred Tales 1) 38, with Festugiere 1986, ad loc. 



332 CHAPTER FOUR 

she had recovered herself she tried, still trembling, to wake me, and 
finding that the fever had left me and that much sweat was pouring off 
me, she did reverence to the manifestation of the god (tt|v |iev tot) 6eot) 
7tpoceKwr|c>ev e7tupdv£tav. . ., 1. 13 If), and then wiped me and made me 
more calm. When I spoke to her she wished to declare the virtue of the 
god (tt|V -cot) 0eot) . . . apernv, 1. 135-7), but I, anticipating her, told her 
all myself; for everything that she saw in the vision had appeared to me 
in my dreams (oca yap 8 id Tn<; oyeax; ei8ev, tatka eycb 8i' ovetpdtcov 
ecpavtaaicoSriv, 1. 138-40). 

POxy 1381, 11. 91-140, tr. Grenfell and Hunt = Totti no. 15 

This sophisticated reflected I-narrative uses the device of an extra- 
diegetic stance within the frame of a homo-diegetic one in order to lend 
the account greater authenticity. 512 What would have been a relatively 
unconvincing narrative: "I saw in a dream..." has been lent author- 
ity by having the events described by someone other than the main 
narrator, namely the mother in a directly-experienced utterance. This 
confirmation is then undermined by the expression (3A,e7tovxeq nev otjk 
oncptpcoc; (itself a trope in this genre), which suggests that, while she could 
not really confirm it, it was after all her piety that induced her not to 
stare rudely 313 This concession is then countered by the dream-vision, 
normally weak, which now appears paradoxically as a sort of by-miracle, 
a divine triumph. What might have appeared as one of those "experi- 
ences insituables que constitue le passe onirique" has become "le passe 
vraiment vecu". 514 Personal appearances by divinity gave a special status 
to the recipient of the miracle, and underscored the temple's claim — 
I take it the experience took place, or was claimed to have taken place, 
in a temple — to being a privileged point of mediation, at a time when 
their economic status was already under threat. 515 



512 To use the terms introduced by G. Genette, Narrative Discourse (Ithaca and Oxford 
1980); cf. M. Martinez and M. Scheffel, Einfiihrung in die Erzahltheorie 3 (Munich 2002) 
47-67. 

513 Similar details are found in other traditions, for example the accounts of the 
appearances of the Lord of the Wounds in Oaxaca in summer 1911: "Most of 
them . . . did not dare to look up at the Lord as the entire house shook": E. Wright-Rios, 
Envisioning Mexico's Catholic Resurgence: The Virgin of Solitude and the Talking 
Christ of Tlacoxcalco 1908-1924, Past & Present 195 (2007) 197-239 at 224. 

511 G. Genette, Silences de Flaubert, in Figures 1 (Paris 1966) 223-43 at 233. 

515 Sauneron 1959, 51 has drawn attention to a Demotic papyrus describing an 
incubation that takes place in a humble dwelling lit by an oil lamp; as incense burns, 
the subject gazes at the flame until the god appears, and then lies down on a mat 
rush to receive the divine message. This is typical of the instructions for divination in 
the magical papyri. 



THE RITUAL SYSTEMS 333 

On the other hand, it has recently been stressed that there is no 
positive evidence in the west for the practice of therapeutic incubation 
in the Egyptian cults: there are no allusions in the Latin epigraphy to 
healing personnel, and it is hard to find areas of excavated temples 
that might have been suitable for regular incubation (Renberg 2006, 
1 14-6). Moreover, Horus-stelai are unknown outside Egypt, suggesting 
that there were strict limits to Graeco-Roman willingness to accept 
certain medical practices. It seems likely that, just as the practice of 
Egyptian 'magical' medicine failed to make headway outside Egypt, 
so Greek incubation failed to make any impact within the cult in the 
Latin-speaking west. Reception of the oriental cults, in this as in other 
areas, was always highly selective. 

This leads us on to the other type of incubation, oracular, or, more 
loosely, to the general subject of visions and dreams in the cult. Per- 
sonal contact with the deity sometimes went beyond inner communion 
and took on an external or apparitional form. This was primarily, 
as I have remarked, a means of registering the special status of the 
recipient: visions and dreams are a constant theme of Apuleius, Met. 
11.19-30, where they chiefly function as a stimulus to initiation. Such 
manifestations might take any of a wide variety of forms. Sometimes 
the god's will was conveyed through miracles worked by the statue, or 
by means of an epiphany of some sort, usually at night. There are 
many epigraphic examples of dedications made as a result of visions 
and visitations, nocturnal and other, icon:' ovap and Kcxxa 7ip6oxay|a,a. >lfa 
Frequently the significance of the response was not sufficiently clear to 
enable a decision to be made about how to act; then one had to consult 
an oveipoKpixrn; for advice. 517 Aelius Aristides provides an example when 
he describes a vision, ovk ovap aXX' vnap, in response to a request, 
in which Apollo/Asclepius/Serapis had held up his fingers in such a 
way as to suggest that he would live either another 10+3 years, or 



516 E.g. IG X.2, 91 = R1CIS 113/0569 (Thessalonike); L. Robert, Hellemca 1 (1940) 
66f. no. 1 = RICIS 112/0705 (Demetrias); IG XI.4, 1247 1. 4 = RICIS 202/0124 
(Serapeum A, Delos). For a full list, see RICIS p. 790, Index 5 s.v. ~. 

517 oveipoKpi-ms: ID 2071 = 202/0217 (Serapeum C, Delos, shortly after 166 BC); 
ID 2120 = 202/0245 (ibid., 129/8 BC); ID 2151 - RICIS 202/0289 (ibid., 114/3 BC) 
etc.; female interpreters (nom. sing.: oveipOKprcu;): IG II/IIF 4771 = RICIS 101/0221 
(Asklepieion, Athens, c. AD 120); ID 2619 - RICIS 202/0209 col. i, 10 (Serapeum C, 
Delos, 95/4 BC); Kpivwv xa opdnaxa: SEG 42 (1992) 157 = RICIS 101/0206 1. 16f 
(Serapeion, Athens, between 116/5-95/4 BC); cf Baslez 1977, 235f. 



334 CHAPTER FOUR 

15 + 2. sl8 Such uncertainties must have been rife when the visitation 
or sign was not specifically requested but resulted from the unilateral, 
and thus unanticipated, decision of the deity. 319 However this office is 
unattested in the Latin-speaking west, and we must assume that the 
institution of dream-interpretation, like that of therapeutic incubation, 
was not received there. 320 

At least in the eastern Mediterranean, nocturnal visitations were the 
Egyptian gods', and especially Serapis', regular means of communicat- 
ing the demand to erect a new statue or construct a new sanctuary. 
Here history and policy came together. Dreams were already a feature 
of the cult of Osarapis at Memphis in the Saitic period; and where 
conventional civic imperatives fail, wonders must supply the deficit. 
One example is the dream in which Ptolemy I Soter saw the colos- 
sus of Pluto at Sinope, ovk zniaxa^evoq o\)S' ecopcxicroc; rcpoxepov oioc; 
< fyv > xf|V nop(pr|v, 'even though he was completely unfamiliar with its 
appearance'. 521 According to Plutarch, he finally found a man named 
Sosibios who told him what it was he had seen — at any rate, he identified 
the king's description with a statue he claimed to have seen at Sinope. 
In a similar vein, Isis is supposed to have told Seleucus IV Philopator 
by means of a dream to transport her 'horned statue' (to poviKepcov 
6cycxA,|j.cx) from Memphis to Antioch on the Orontes. 522 Such messages are 
a stock-in-trade of the aretalogy The Delian Serapis-aretalogy recounts 
how the god appeared in a dream to Apollonius, the grandson of the 



318 Aelius Aristeides, 0x48 (— Sacred Tales 2) 18, with Festugiere 1986, ad loc. 

,19 Sometimes appearances of Serapis were less encouraging: Artemidorus, Oneir